The Art of Language Invention: a review

                I received an advance copy of David J Peterson’s The Art of Language Invention, which after some procrastination I have recently finished. It’s a book I highly recommend both to newbie and advanced conlangers, as well as anyone who might be interested in conlanging. Before I get started with my review, I will point out that William Annis has a great overview of the book on his Tumblr, and Gretchen McColloch of All Things Linguistic livetweeted her reading, and her take, as a non-conlanger, is really fun.

                David is very good at introducing basic concepts of linguistics in an entertaining and understandable way. There are, of course, those points that make me go, “Ah, yes, that’s the simplified lie we tell the undergrads,” but that part of the book isn’t for me, and I highly suggest anyone interested in conlanging to read the book especially if you don’t know anything about linguistics. I grew up in conlanging with the web version of The Language Construction Kit (years before it became a book), and I wish I’d had this book as well. It really is a beautifully laid out and easy-to-digest.

                All that said, as a linguist and a moderately skilled conlanger, the most valuable part of the book was in the copious examples and especially the case studies provided in the book. I was interested to see how David’s approach differs from mine, and what I could learn from him. Some of these things are just a function of different experiences with language. For instance, in his case study on Irathient, he discusses how he wanted to make the language “slow”, and that his prototype for a language with a slow speaking rate was Inuktitut, a language with a massive derivational system that packs a large amount of meaning into a word via derivation. Irathient is by no means like Inuktitut – morphologically it has more in common with Bantu languages – but I wonder if I might have approached that problem differently. The reason being, my idea of a prototypical “slow” language is Mandarin Chinese – almost the opposite of Inuktitut in that it is an analytic language where almost every morpheme is a single, complete syllable, and the majority of derivation is in the form of two-syllable compounds. In addition, whereas David feels the need to build in agreement in case lines get cut down in editing, my experience with Mandarin would make me comfortable with a language where you have no agreement but can find missing elements by context. Neither of these approaches are right or wrong, and I think increasing one’s repertoire of tricks and language structures can only be good for a conlanger.

                Another side of it is David’s focus on lexicon building and historical derivation. This is a place where I have to say David is far better than me. I’ve only recently started building a conlang from a lexicon-centric position, and seeing how David builds his words is very helpful. Look out for his example of an entry in his Sondiv dictionary, which already surpasses any entries I’ve made in a conlang dictionary for completeness and number of terms (of course, it is an entry for a triconsonantal root, but I think any derivation system should be built in a way that can handle this complexity).

                In the end of the book, David speaks briefly on the status of conlanging as an art form, and also on how an economy for professional conlanging can evolve. David encourages authors of speculative fiction to collaborate with conlangers, even if all they can offer is a percentage of royalties or the conlanger’s name on the front cover. I like this idea. I’d be more than willing, given the opportunity and the time (oh where to find time!), to collaborate with an author in that way myself, and I think there are a lot of very good conlangers who would as well. I would like to add, though, that even if you are a conlanger yourself and writing some creative work, I think that collaborating with other conlangers can be a benefit. Lots of conlangers have skills and knowledge applicable to a particular type of conlang (e.g. non-humanoid alien languages, or historical a posteriori languages). But definitely, definitely, I wholeheartedly agree that creators should partner with conlangers.

                The Art of Language Invention is not a comprehensive guide. If you are a beginning conlanger, this book is your starting point. As McCullogh put it, it’s “a geek’s guide to linguistics”, and something that makes for a good introductory text. If you are an experienced conlanger (or an intermediate one like me), it’s a window into how another conlanger does his craft, and being exposed to different approaches can only be beneficial to your own work. All in all, I recommend it to anyone interested in our weird little hobby.

Lexember 2014 #7: brizit

Another day, another word:

brizit vI to be stubborn

I've decided that at least most verbs in Middle Pahran will behave as Class I verbs, basically stative verbs as many languages have. Another small note, brizit does have a superficial similarity with briiza "donkey", however, I don't think I will jump through any hoops to make them related. Brizit is just what the random number  generator in awkwords gave me today (well, the proto-form bridit, which I applied sound changes to), and I think I'll just let it remain a coincidence.

Lexember 2014 #6: bun

As I said yesterday, I decided to limit the semantics of batlaam to only include larger rivers, simply out of personal preference. So, today, I decided to make another term for the semantic space that's left by that.

bun ni stream, creek, tributary

As suggested by "stream" and "creek", bun is usually for much smaller bodies of water than batlaam. However, in the sense "tributary", it can be quite large. Thus, if it is used to describe a body of flowing water in isolation, that's likely a small, easily fordable stream, but if it is referring a body of water that flows into another, it could be quite large.

Sometime when I'm not fiendishly busy with other things, I will work out some examples.

Lexember 2014 #5: batlaam

A new day, a new word. Here is a word I was surprised I hadn't had yet:

batlaam ni water, river

Once again, am using an asssociation I found in A Conlanger's Thesaurus. However, the next couple of days I may refine this a bit with a couple other words. Hearing batlaam, I kind of want it only to apply to big rivers, smaller streams will have another word. I'm also thinking that I will have another term for water, as I eventually want there to be one term that becomes a technical term used in alchemy, while another is the common term (Pahran is for a fantasy world).

Also vacillating on whether this should be animate or inanimate, since a river could be conceived of as animate. Thoughts?

Lexember 2014 #4: tutii

It looks like I'm a tad late again today. Too much to do all at once, but I wanted to get my Lexember entry in.

tutii ni grass, wheat, field

This word may end up being revised later, as I haven't decided what Pahran food culture is like and whether it will rely all that significantly on wheat (which might affect the likelihood of a wheat>field relationship), but I will note here that when I'm equating grass and wheat, note that Pahrans don't have cut grass lawns, they'd usually encounter long grasses. The relationship between grass and wheat is clearer to them than it might be to say, a 21st century American.

Lexember 2014 #3: 'ããs

Today's word is my first verb for Lexember:

'ããs vII to train (an animal)

'ããs is a class II verb, which is the default class for transitive verbs. I won't go into the morphology, as I think as soon as I have time to work on Pahran again I may be revising the class II verb paradigm.

However, the most interesting thing about thinking of 'ããs today is that it led me to revise a completely unrelated word. Previously frujmaa "to teach" had also been class II, but I thought it might be better to keep 'ããs as class II and move frujmaa to class VII, which typically indicates situations where the subject is the origin or creator of the object (verbs such as build, create, write, etc). I also created a separate entry frujmaa "to learn" which is a class I (intransitive) verb.

My idea is that the class VII frujmaa "teach" is derived from class I frujmaa "learn", and that frujmaa will retain class VII status in contrast to 'ããs because of the different animacy relations. That is 'ããs is only used for animals, and thus only occurs when the object has lower animacy, whereas frujmaa applies when teacher and learner have equal status.

In any case, it looks like I am submitting this just under the wire. Hopefully I'll have the time to do the next few words in more timely fasion :P

Lexember 2014 #2: sõõp

It seems that I'm just a tad late for this (by my own time zone) but my second entry for Lexember 2014:

sõõp ni bowl, cup

My inspiration for the meaning of this word by my recent watching of lots of Chinese historical dramas (at the behest of my wife, of course). One thing you will learn about ancient China through these dramas is that apparently everyone used to drink alcohol out of bowls. I haven't done so much research into this, so I'm not sure why ancient Chinese are depicted this way or how accurate it is, but it does make sense that a cup and a bowl can function similarly and really are similar implements. As such, I decided that Pahrans, in the time period when Middle Pahran is spoken, will be sort of in between using just bowls to using more cup-like containers for some things. Eventually sõõp may come to mean just "cup" in a daughter language, or it may differ in meaning between a couple different dialects/daughters.

Lexember 2014 #1: tuuzaa

If you don't know, Lexember is an even that conlangers have latched onto where we invent a word in one of our conlangs for every day for the month of December. Until now I've never really participated, but I thought I'd go ahead and go for it, since it only takes a few minutes out of my day. So here it is, Lexember 2014, Day 1, from Middle Pahran, my current (languishing) project:

tuuzaa ni roof, house, bedroom.

I got the association of "roof>house" from one of the semantic maps from William Annis's A Conlanger's Thesaurus, which I think anyone doing Lexember should look to frequently for inspiration. This particular association works well for the fictional culture I have in my mind for Middle Pahran. Pahra is meant to be in a tropical location, and homes are fairly open. Poorer homes may be little more than a thatch roof, but even richer people with sturdier homes will culturally prefer entertaining guests or relaxing in the saamas, a central courtyard or flower garden, whenever the weather permits, and might not want to spend so much time indoors. Hence, I get house>bedroom (somewhat close to house>room on the map). In a Pahran inn, each room would be called a tuuzaa and would face directly onto the courtyard with no interior hallways connecting to other guestrooms.

Hopefully I can keep this up amid the craziness of papers and whatnot. In the mean time, have a great Lexember everyone!


I've been using the online memorization tool Memrise to review my Chinese. Memrise uses a spaced repetition method, presenting vocabulary items at different times according to how many times the learner has succeeded remembering them, similar in some ways to Duolinguo and the desktop app Anki. They add another element to these systems in the form of "Mems", or customizable visual aids. These can be useful when learning basic vocabulary, though I haven't made so much use of it recently.

Memrise has the advantage of many crowd-sourced lessons, and the ability to create and show your own lesson material. This, of course, means that finding lessons means wading through potentially hundreds of submissions for more popular languages, with varying quality (some of these lessons have some video instruction built in, while others are only simple text). So far, I have spent the most time with a vocabulary set based on the HSK (汉语水平考试 hanyu shuiping kaoshi), but there are a variety of lessons available, and the site lists over 130 languages, including a good number of African and Native American languages, eight sign languages, and even a number of invented languages (including Esperanto, Na'vi and Toki Pona), as well as a host of other topics you can review through the site. Memorization sessions are fun and engaging (if occasionally frustrating when I have trouble with a word).

No tool can replace speaking practice for learning a language, and as said before, the user-made lessons are bound to vary in quality, but all in all I see Memrise being a good place to look if you want some supplemental vocabulary building practice to go along with your overall language learning. And it's currently entirely free, which is always a plus.

Drawing from an urn

Today's XKCD had a joke with an interesting linguistic angle:

I think this kind of humor illustrates two important things about words: First, you can't separate a word from it's associations and connotations. Second, those associations are different for different people in different situations.

For most English speakers, an urn is primarily a container for the ashes of a deceased individual, but for this teacher, in a discussion of statistics, it's just a container for randomized balls in an urn problem. But it's not a pun and these aren't homophones; both of these people are probably picturing a ceramic jar of some kind, they have the same core meaning for the word. It's just the clash of associations that makes the joke.

Language is a way of describing the world, and our understanding of the world is always affecting how we interpret it.


Adventures in Linguistics: He is X nor Y

I had a curious experience in semantics class today.  We were covering the scope of negation, and the professor had presented us with three sentences:

(1) Pat isn't a plumber and isn't an architect.

(2) Pat is not a plumber or an architect.

(3) Pat is neither a plumber nor an architect.

Part of what we were discussing was the fact that all three of these sentences mean the same thing  (that is, the sentence is true only if Pat does not belong to either of these professions), but it seems that (1) and (2) derive that meaning differently, and we were working on which of those sets of rules apply to (3).

I won't bore people with the technical details, but along the discussion, one of my classmates brought up an example of their own:

(4) Pat is neither a plumber or an architect.

Which was grammatical to her, though I find it slightly questionable.  This encouraged me to bring up an example that I had been mulling over in my head for about 10 minutes:

(5) Pat is a plumber nor an architect.

Though I thought that (5) was good and means the same as (3), apparently no other native English speaker in the class agreed with me that (5) was grammatical at all.  One person thought it may have to do with me being from "the South" -- which still amuses me, since I never did consider the part of West Virginia I come from particularly Southern (I suppose it looks very different from Wisconsin).  In any case, it did lead to a short discussion of what could possibly be going on with my dialect of English to cause this construction.

It's funny how these things pop up.  I've had a moment like this before, when the double-modal might could was brought up in syntax class (that one I know is common in Appalachia and the South, but not up here), and I'm sure these things will happen again.

Short Stories blog

I've added a new blog where I'll try post up short stories from time to time.  I really would like to make a habit of writing more often.  Occasionally, stories just pop into my head or are inspired by a dream and I have to get them down, but I don't really write as often as I'd like.  Anyway, the first entry is a very short post-apocalyptic horror vignette inspired by a rather Cthulian dream I had.  Maybe someone will like it.

From Aeruyo to Malviz: Where is this phonology even going?

So, one issue I realized I would have to deal with in deriving a language is historically is the fact that I would have to go back and analyze the phonology to figure out what the heck phonemes were there, anyway.  So, after a lot of wrangling, I managed to get Zounds to apply my changes to my entire lexicon.  Now after doing that, I'm finding the analysis will be a daunting task, and I'm feeling lazy.  So I thought, why not show the word list to my conlanger friends and see what they think of it.

So, with no explanation, I have here a list of all the words, in phonetic transcription, without the original Aeruyo words.  I won't tell you anything about what I've worked out from my initial eyeballing to keep it pure.  So, if you feel like doing some phonological analysis: here is the word list.

I'll come back when I've had time and desire to work up my own analysis, as well as any tweaks I've made.

From Aeruyo to Malviz: A little morphology

Since my last post on deriving Malviz from my existing Malviz language, I've worked a little on the morphology of the language.  Aeruyo had a complex inflectional system on both nouns and verbs, so in order to work out the morphology of Malviz, I simply ran several fully declined nouns and several fully conjugated verbs through the sound changes to see what irregularities and mergers occurred naturally.  Then, based on this work, I made the following morphology changes:


  • I got rid of the vocative case.  In most cases it was merging with either nominative or instrumental, and I had planned on getting rid of it anyway, so it seemed like a good opportunity.
  • I merged the plural and collective for spirit nouns, since apocope had essentially done that for me right out of the gate.  In cases where the collective form causes o>u mutation (particularly in oral ~ uro > oral ~ urz), I kept the mutated form as the plural/collective form, though analogical flattening in some roots isn't ruled out.
  • All verbs had the non-past positive and negative forms merging, while the past forms remained distinct, so I extended the past tense negative forms to cover non-past as well (essentially creating a tenseless negative form and avoiding the need to create a negative particle.


There are still a few cases where forms are identical, but for the most part those are quite regular.  One issue I have with verbs is that they are merging in different and interesting ways depending on the root, which is making it hard for me to decide what forms to keep.  For instance, often the potential and optative moods are merging in both positive forms but not the negative:

Indicative Potential Subjunctive Optative Necessitive
Past khon khonrz khonm khonrz khoŋrz
Non-Past khonv khonrz khonai khonrz khoŋrz
Negative khongui khonrui khonmmoi khonzui khoŋgrui

However, there are cases where it does not merge:

Indicative Potential Subjunctive Optative Necessitive
Past adeh adez adem aderz adegrz
Non-Past adev adez adei aderz adegrz
Negative adekui aderui ademoi adezui adegrui

And there is one rare case where the necessitive also merges in with potential and optative (again, only in positive forms):

Indicative Potential Subjunctive Optative Necessitive
Past per perrz perm perrz perrz
Non-Past perv perrz peroi perrz perrz
Negative pergui perrui permoi perzui pergrui

I probably will apply some sort of analogical flattening for the last case, since it requires such a specific initial configuration (an Aeruyo verb root CVrV-), though the more common merger of potential and optative is quite interesting.  Should I just completely merge one to the other (the pronunciations are quite close, after all), or should I say, keep the distinct optative negative form for some vestigial usage?

From Aeruyo to Malviz: Starting with Sound Changes

It's been a while since I did any significant conlanging, so I thought I'd share some of my most recent efforts.  Some people may be familiar with Aeruyo, which has a grammar posted on this site.  Within the same world that Aeruyo and its speakers, the etherial Aeruro, exist, there are also the Malviz.  The Malviz are another group of spiritual beings who split off from the Aeruyo in time immemorial and cover and are essentially the "dark" version of the air spirits.

Malviz speak a decendant of Aeruyo, the conceit being that Aeruyo did not actually change much because its primary speakers are immortal and have separated themselves more from the physical realm, whereas the language of the Malviz has changed slowly but surely due to their constant interaction with the changing world through possession of undead.  This may be a very flimsy hand-wave (and may need beefed up in my stories), but it allows me a nice sandbox to play with historical changes before I get serious about working out the human languages of my world.

My process for making the sound changes to Aeruyo to Malviz went somewhat backward.  I had a couple of names that I wanted to fit into the ending Malviz language -- namely Kavrz [kʰavʐ]* "Malviz incarnate of wrath" < kafira "anger" and Malviz [malvɪz] < malefiri.  After building the sound changes that would result in those two forms, I built out a couple more changes.  Here's what I came up with:

  • V > 0 / _#
  • stress shift to first syllable
  • a, e, o, u > ə / [-stress]
  • i > ɪ / [-stress]
  • [-aspirated] > [+voice] / V_
  • j, w > 0 / [-continuant]_
  • ɾ > ɻ
  • V > 0 / [+stress]._
  • ɻ > ʐ / _#
  • [+aspirated] > [-aspirated] / ._[-stress]
  • w̥ > ɸ

There are no strict time frames here -- again, I am building these languages kind of in a sandbox, taking advantage of the conceit that they are spoken by immortal spirits who reject influence of mortals, etc etc.  I may add a few sound canges (I'm looking at diphthongization) or rejigger the order, but so far this seems to be a good start for me.  I feel the next step is to use these and run my inflectional paradigms through Zounds and then work out what additional morphology changes follow from that.  I already know that I'll be losing the negative verb forms to that very first apocope, so I'll need to make a negative particle -- I plan on using men "never".


*Yes, these are phonetic transcriptions.  I will have to work out allophony after I have figured out precisely how the sound changes are affecting everything.

Conlang Language Options in Minecraft?

While looking over the patch notes for Minecraft 1.2.4, I noticed a section under the known bugs labelled "Translation Related".  There, in addition to a lot of notes about Spanish translations that mostly seemed to involve correcting names (including some interesting juggling of the terms castellano and español that might be deserving of its own post), I found this curious and rather amusing line:

The translation [Quenya (Arda)] has "Lever" labeled as "Mechanic Pen*s"

A quick check reveals that Minecraft is actually available in three constructed languages: Esperanto [listed as "Esperanto (Mondo)"], Quenya ["Quenya (Arda)"], Klingon ["tlhIngan Hol (US)"] ...  Why Klingon's listing is US and not some term for the Klingon Empire or their homeworld Kronos/Qo'noS I wouldn't know.

The trivia on Minepedia's Language* page does not redact the term, so I presume that some joker did indeed name the Lever element "Mechanical Penis" (Minecraft uses a crowdsourcing site for translations, and it has gotten them in bigger trouble than this.), however, the problem was apparently fixed, as when I jumped in the game using the Quenya UI and made a lever, the mouseover text read "Turolwen" as shown in the image below.

I can't vouch for the accuracy of any of these translations of course, though the Quenya is obviously incomplete, as a few English words and phrases are still being used.  Of course, I'm sure that many of the words Minecraft needs would not be in any canonical Tolkien source, and I think the Elven language people tend to be a little touchy about coinages -- it's just one of the things that can get them arguing.

In any case, it's cool to see people having fun with some conlangs.  In addition to the proper conlangs listed above, there is also a hilarious joke language called Pirate English in the options, and it's pretty much exactly what you would expect it to be.  And of course, there are a wide array of natural languages, too, which will of couse benefit Minecraft a bit more.

*Which, as I write this, does not list Esperanto, though I'm sure that will be corrected.

How is Huntsman's China experience a bad thing?

Have you seen this monstrosity?

This ad makes me angry.  It's not because I support Huntsman in any way, while to my mind he's a better candidate than the other Republican candidates, no one in that field interests me (and unfortunately, I only see Barak Obama as marginally better).  No, it's the fact that it takes a number of multicultural and international appeals of Huntsman: bilingualism, adopted children from China and India, a deep understanding of China -- and casts these qualities that I think would be great in a President, and presents them as bad or evil.

Know upfront that I won't scream at Ron Paul for this.  Though this is my first time seeing the actual ad, I had heard about the controversy and the story that it was a supporter of Paul's who created the ad, unknown to him, and that Paul disavowed him.  I have other reasons for being uninterested in Paul, but so far I have no information that would contradict those statements.

What I am angry about is that whoever created this ad apparently thinks that bilingualism and international experience are bad things to have in a president, and that same person would also exploit two little girls to prove his point.  In what world is that OK?  Really, in what world does that even make sense?

We live in a global economy, and in a world where interacting with people accross the globe is a necessity if we are to succeed.  I want the President of the United States to speak Chinese, Spanish, Arabic, Russian, French ... as many major and widely used languages as possible.  I want a president with a wide range of international experience, who has studied abroad, worked abroad, and lived abroad.  All of this will facilitate communication and understanding when the president is negotiating with foreign governments.  Yes, I want him to be furthering American interests, but I want him to have cultural and practical knowledge that will help him in doing that.

There is no reason that someone's ability to speak a foreign language or their experience in a foreign country (barring them working for that country, which Huntsman wasn't -- he was a student and then the US Ambassador -- working for our country) should be seen as anything other than a positive in terms of one's qualifications to be President of the United States.  We need skills and experience like that in our top offices.  And if we universally rejected people with those qualities in the highest positions in the country, we would not have risen as the most powerful country in the world.

EDIT: It's a good thing that Huntsman knows how his Mandarin skills should be viewed: Judging from how he used them on the debate floor.  Pull out a chengyu next time, sir!

Moving Domains

I have just moved from GoDaddy to  I realize I'm a little late in joining the boycott, but I finally found a time to get it all worked out.  I've been thinking about leaving GoDaddy for some time ( was on Hover from the start, specifically because I was annoyed with GoDaddy), but their support of the Internet-breaking leglislation in SOPA and PIPA was the straw that broke the camel's back.

There shouldn't be any effect on visitors to the site from the change of registrar.  I just felt that the need to share my reasons.

From My Conlanging Past

I was cleaning out my room today and found an old binder done up as a "spellbook" in Aerol (the predecessor to my constructed language Aeruyo -- which I am in the process of putting finishing touches on a grammar for).  It's been so long now that I have trouble deciphering the old Aerol script.  It doesn't help that I created a horrific featural monstrosity that I hope no one with dislexia would attempt to learn -- literally distinguishing characters by rotation.

But anyway, I thought I'd share some images:


I believe this cover reads "Sagal tan Hatal", roughly "Summonings and Wishes" or some such nonsense.  The small text might mean "Written in the Aerol Language, by Fondor", Fondor being a pseudonym I used to use (what's here is the Aerol reflex of "Fondor", of course).  The following are the spells I wrote that for the life of me I cannot read right now, and I don't feel like taking the time to decipher them (I lost the key to this script a long time ago, and it will take some time to figure it out.