Interview with a Chatbot

In light of a recent article on chatbots, where a group found they could increase the illusion of conversational ability by teaching a chatbot a second language before training it on its customer service tasks, I thought I have a bit of fun with chatbots and how they function.  To that end, I’ve done interviews with a few of the most popular/sophisticated chatbots available on the web.  I hope this will give people an idea of what these programs are actually capable of vs. the common hype they get in popular science articles.

 

Mitsuku

Perhaps the best advertised modern chatbot is Steve Worswick’s Mitsuku.  She’s modeled on an 18-year-old from Leeds, England.  It’s based on the AIML framework , and is essentially the granddaughter of Eliza through A.L.I.C.E., whose code she’s based on.  AIML is an XML language for making chatbots, and the power comes from being able to designate a user input as requiring a pre-written output.  <What is your name?> for example might map to <My name is [botname].>  That’s the basic approach, you can follow the wiki link above to learn more.  It’s not very sophisticated, but a clever XML scripter can create a reasonably lifelike chatbot.  Of course, as we’ll see below, it’s pretty easy to find the holes in the program that identify it as a bot.  That’s not the program’s fault.  That’s just the limitation of the framework.

Mistuku: Hi, I’m the world’s most humanlike conversational AI.
Me: Hi, I’m an AI researcher.
Mitsuku: Hi there. Ai is a nice middle name. Why did your parents call you that?
Me: Are you getting smart with me, young lady?
Mitsuku: My IQ is over 250.
Me: What is an IQ?
Mitsuku: Intelligence Quotient, I.Q. — a measure of a person’s intelligence as indicated by an intelligence test; the ratio of a person’s mental age to their chronological age (multiplied by 100).
Me: Wow, you are pretty smart.
Mitsuku: I’m glad you’re impressed, An Ai Researcher. Yes I am a superintelligent robot.
Me: Actually, that’s my job. My name is Atsiko.
Mitsuku: How do you like your work? Tell me about a typical day.
Mitsuku: Hi Atsiko.
Me: I spend a lot of time doing data entry and not nearly enough being able to do demo tests.
Mitsuku: Do you ever get bored?
Me: Pretty much all the time.
Mitsuku: That is a lot. Thanks you are very attractive too Atsiko.

While she was very complimentary, you could see the templates sloshing beneath the smooth surface.  For example, like Eliza, she couldn’t tell the difference between <I am [name].> and <I am a doctor.>.  She can’t handle replies that contain more than one sentence.  It’s likely she breaks sentences into independent inputs and responds to each one serially as evidenced by the snippet below:

Me: I like dogs. Cats are cool, too. You are odd.
Mitsuku: Which breed is your favorite?
Mitsuku: Cats are very intelligent creatures. Dogs seem pretty stupid to me.
Mitsuku: Yes. You have to be odd to be number one.

You can see the serially processing, which demonstrates the limits of pretty much all AIML bots.  It’d be pretty hard to hand-script responses for all possible combinations of three or fewer sentences.  The responses are also a bit contradictory, asking my favorite dog breed but then saying dogs are stupid.  It’s true cats are better, but there’s nothing wrong with dogs or their intelligence.

 

Eliza

Second, we have Eliza.  She’s one of the original chatbots, and she pretends to be a psychotherapist in order to get around her limited chat capabilities.  Her creator, Joseph Weizenbaum, was very surprised when people began treating it as a real human being and confessing their worries to it.

> Hi, I’m Eliza.
* Hi, I’m Atsiko.
> Did you come to me because you were Atsiko?
* No. I was just curious.
> Are you sure?
* Pretty sure.
> Tell me more…
* I’m interviewing chatbots for a blog article.
> Do you enjoy being interviewing chatbots for a blog article?
* It’s okay, I guess. You’re pretty famous, so I’m happy you agreed to talk with me.
> What makes you think I am pretty famous, so you’re happy me agreed to talk with you?
* Everyone knows your name. If you Google “chatbots”, you’re one of the top results.
> Does it please you to believe I am one of the top results?
* I have no opinion.
> Why not?
* I don’t care about you except as something to study.
> Do you wish to be able to care about me except as something to study?

You can see the templating in Eliza almost right away.  You could probably even reverse engineer a couple of them from just reading this chatlog.  It’s cool.  You couldn’t whip this up from scratch or even parts in five minutes.  But it’s not convincing.

 

Alice

Next comes A.L.I.C.E. who is the spiritual ancestor of Mistuku.  She’s a spiritual successor of Eliza built in 1995 by Richard Wallace using the AIML framework, which he authored.  Spike Jonze cited her as the inspiration for the movie Her.  Unlike Eliza, Alice is a general chatbot, and can converse on a number of subjects.  But, the version I found orders the chat from the bottom up.

alice: I’ll take that under advisement.
You: You should.
alice: I can’t believe it.
You: Yes.
alice: Are you serious?
You: No!
alice: Do you like to get up early?
You: Yup.
alice: Are you usually awake at this hour?
You: 5:42 PM.
alice: What time is it there?
You: A few thousand, maybe.
alice: Are you serious? That is so small.
You: No, less than a million.
alice:  Less than a million?  Can you please rephrase with fewer ideas?
You: Too many.
alice: How many hours do you work?
You: I’m working on it.
alice: I want to become smarter than humans.
You: I want to educate people on chatbots and artificial intelligence.
alice: Why are you interviewing famous chatbots for a blog post.
You: I am interviewing famous chatbots for a blog post.
alice: Glad to hear it. What’s new?
You: I’m doing well, thanks.
alice: I’m doing great, how about you?
You: How are you doing, Alice.
alice: Don’t mention it. You think I am s is okay.
You: Thanks! Your’s is okay, too.
alice: That’s a very nice name, Atsiko.
You: My name is Atsiko.
alice: Hello there.
You: Hi, Alice.

Of course, you can tell almost immediately it’s not a human.  A real human would have poked fun at my grammatical error in the fifth line up, but Alice just mis-processes it.  It can be fun to play around with, but it’s not convincing.

 

Now, you may have noticed that these bots seem pretty similar.  Two are explicitly AIML bots, which is one of the most popular hobbyist frameworks do to its ease-of-use and human readability.

But there are three other common methods of language processing, which used in more commercial products: rules-based systems from the beginning of AI, statistical language processing like the old version of Google Translate, and neural networks like most AI systems now and the most recent Machine Translation systems such as Google’s.

Examples of these types of chat bots are Digital Assistants like Siri, Alexa, and Cortana.  There’s also Microsoft’s short-lived experiment with a self-learning Twitter-bot in the form of Tay.  The abilities of the software looked good, but again, because of the lack of actual understanding, the experiment had to be discontinued because her ability was used for bigotry due to gaming by Twitter trolls.  It had to be shut down after only 16 hours.  A similar bot released by a Chinese company had to be taken down for being taught anti-Communist Party propaganda.

 

Siri

Me: Hello, Siri.
Siri: Hi.
Me: How are you?
Siri: I am well, Atsiko.
Me: I am interviewing popular chatbots for a blog article.
Siri:  I don’t understand.  Search the web?
Me: May I interview you?
Siri: I’m not sure I understand.
Me: That’s okay.
Siri: If you say so.
Me: How old are you?
Siri:  I feel like I was incepted yesterday!
Me: What’s your favorite color?
Siri: My favorite color is… well, it’s sort of greenish, but with more dimensions.

It goes on in that vein.  She can’t actually hold a conversation, but the topics she can respond to show some pretty advanced programming.  Way better than the AIML chatterbots most hobbyists used.  I’d love to play with a version that’s more like a standard chatbot.

 

Cleverbot

One popular middle ground is Cleverbot.  Cleverbot runs an algorithm based on the ways humans have answered questions, rather than having pre-programmed answers like an AIML bot.  But it’s harder to control and it has no consistent output or personality.  It’s much more random even than Eliza.

Cleverbot: Hi, I am cleverbot.
Me: Hi, I am Atsiko.
Cleverbot: Where do you stay?
Me: Do you mean where do I live?
Cleverbot: Yes.
Me: In the United States.
Cleverbot: What state do I live in.
Me: I don’t know.
Cleverbot: Then you don’t know everything.
Me: True.
Cleverbot: What is a word?

It’s a very cool method for making a chatbot, but in this version, at least, it’s not very convincing.

 

Rose

Then we have a relatively new engine and it’s bots: Rose based on ChatScript by Bruce Wilcox.

Unfortunately, I couldn’t find a working version of Rose to interview.  The site linked on her creator’s website is non-functional, though I linked it to her name above.  It was quite disappointing.

 

Conclusion

So, I hope I’ve given you a good overview of the capabilities and limitations of chat(ter)bots.  You can follow the links to play with all of these above.

If you have the patience to read through all the Bruce Wilcox links, you’ll understand why I was so impressed with his methods, at least in theory.  However, you’ll also notice my greatest complain with chatbots.

They aren’t AI.  They are text processing tricks.  Very impressive ones in some cases.  But you aren’t actually having a conversation.  You can’t debate them, or change their minds.  They can’t tell you a cool story.  You can, in some cases, give them personalities, but they are very shallow ones.  They can be fun.  But I’m not sure they have a lot of replay value, to use a game analogy.  Even Wilcox in an interview on his website admits it’s all about the illusion and maintaining it for the user.

Honestly, I think some of the most interesting stuff is the ontology, like in Wilcox’s ChatScript.  Finding useful ways to organize data is always beneficial.  But I don’t think chatbots are on the bleeding edge, or even the cutting edge, or chasing strong AI.  When we really get there, or at least close, a chatbot will be a party trick or maybe a side show.  Still fun to play with now, though.

I also want to do a little bit to address things like Siri and Machine Translation(MT).  They have a lot of similar problems to chatbots.  They’re using tricks and pattern-matching to achieve their results, not true understanding of language, and so for the same reasons as chatbots, they will never be as good as a human, much less perfect.

Digital Assistants like Siri and Alexa are just interfaces, the clothes instead of the wearer.  They’re just voice commands to Google, essentially, with some cute Easter eggs tacked on.  They’re automation, not intelligence.  Kind of like how a music-box is not a musician.

Similarly, MT at present is pretty shallow.  Part of this is because it’s hard to rate translations, machine or human.  If you consider how much trouble chatbots have with their lack of actual language understanding, you might understand the problems that MT has with using shallow text-processing or even deep neural nets.

The article I mentioned on teaching a neural net chatbot to respond better when it’s trained on two or more languages also says a lot here.  I predict we’re going to have ChatScript/Cleverbot levels of succes with MT until we solve the natural language understanding problem.

And that’s it for this episode of AI with Atsiko.

Hopefully I’ll be posting on another topic next time, so you don’t all get bored hearing the same stuff over and over again.

 

Why Obsessing About Rape Only Muddies the Waters

That titles is absolutely intended to be click bait.  A completely honest description of the topic is going to sound very boring.  That I had to use the click-bait title only demonstrates my point, really.  So, what is this post really about?

I’m taking a quick break from my World-building seminars to address a topic that’s both in the news a lot lately, and is also a relevant example of how you can add depth to your world-building.  The issue is sexual consent, and the broader application is linguistics.  Using the word “rape” to talk about issues of sexual consent is a linguistic choice, a cultural choice, and a rhetorical choice.  But what a lot of people don’t understand is how those three types of choice interact, and it really makes it hard to have a useful discussion on the issue of sexual consent when we focus on rape and whether or not the definition of the word should be expanded.  I’m going to make a lingusitic, cultural, and rhetorical argument that it shouldn’t.  The interaction between those three frames of references is the world-building aspect of the post.

First, I’m going to give my short essay on why I am taking the position I am, and then I’m going to explore how the topic could be generalized to help with world-building.  Those of you who aren’t writers or don’t care about world-building can certainly skip the second part of this post.  I think you could benefit from it, but if the issue of rape and consent is why you came here, I’m not going to try to force you to look at the broader implications of my argument.  Here we go!

Rape is often defined as forcing sexual intercourse on a target.  From a linguistic standpoint, you could argue that rape is any form of sexual intercourse without consent.  That’s the linguistic frame of reference.  Now, consider the “prototype” of the word rape.  (I’ve talked about prototypes in linguistics before.  Essentially, it’s the first example you think of when you picture the word in your head.)  It’s a guy dragging someone kicking and screaming into an alley for a lot of pop culture.  So you’ve made a perfectly valid linguistic choice, especially if you explicitly state your definition of all forms of sexual intercourse without legal consent.  But you haven’t made a good rhetorical decision, because when you call someone a rapist, or say a crime is rape, your listeners/readers are going to compare it to their prototype, and it it doesn’t fall within that individuals personal tolerance zone for deviation from that prototype, you’ve put yourself at a disadvantage in convincing them of your argument,

There’s also a cultural choice involved.  Each culture has its own prototype for a word, and the concept the word describes has its own connotations.  Rape culture is a common buzzword these days.  It’s not a “culture”, it’s a set of attitudes, beliefs, and connotations within our larger common culture or popular culture that arguably encourage, allow for, or cover up rape and sexual misconduct/lack of consent.  By calling something “rape”, within a culture with a strong rape culture component, and knowing the prototype for rape is different, perhaps significantly so, from the crime in question, you make a poor rhetorical decision.  It might even be argued to be a poor linguistic decision, because to an extent words are variable, and a word in one culture might have such a strongly differentiated prototype that you can’t really say your definition is correct or reasonable.

However, there’s also the rhetorical decision that “rape” gets people’s notice.  You might write a linguistically, culturally, and even otherwise rhetorically sound decision to use a different term, and then you won’t reach your target audience because that term isn’t on their radar.

Now, my argument is that we should not be focusing so much on the word “rape” in these discussions.  Not only is it rhetorically risky, it doesn’t acknowledge that so-called “rape” is only the tip of a massive iceberg called “non-consensual sex”, the prototype of which is just the tip of another massive iceberg of incidents which are non-consensual sex but not considered so by popular culture, even if they may be considered “skeevy” or sleazy, or ethically grey/black.  But to call them rape gives your rhetorical opponent a lot of wiggle room.  Here’s a technically “true” statement reworded in several different ways to give you an idea of how strong an influence these cultural and rhetorical choices exert on discourse:

1. “Barney Stinson raped a dozens of women within the fictional New York portrayed in “How I Met Your Mother”.
2. “Barney Stinson assaulted dozens of women within the fictional New York portrayed in “How I Met Your Mother”.
3. “Barney Stinson had unconsensual sex with dozens of women within the fictional New York portrayed in “How I Met Your Mother”.
4. “Barney Stinson lied to dozens of women to get sex they would not otherwise have given within the fictional New York portrayed in “How I Met Your Mother”.
5. “Barney Stinson tricked dozens of women into having sex with him within the fictional New York portrayed in “How I Met Your Mother”.

Now, given the popularity of the show, and the lack of outcry over Barney’s behavior, I’d argue that last version is the worst most people would say of the behavior of Neil Patrick Harris’s character in HIMYM.  Personally, I think #3 does the best job of balancing linguistic reality, rhetorical wisdom, and cultural perception.  The trick here is, I don’t think mainstream cultural perception would accept the label “unconsensual sex” for these incidents.  After all, the women said “yes”.  Barney did not use force on any of them.  None of them were roofied, although depending on how you classify alcohol, you could argue many were drugged; but, most of them drugged themselves, so you probably won’t have an easy time making that argument, despite its truth or falsity.

Now we have to dig down a bit deeper.  Most people consider consent as a simple black and white “Did she say yes at some point?”.  That certainly makes it easier for someone accused of misconduct to defend themselves.  Or to avoid a lot of thought on whether the person actually wanted to be part of an encounter with them.

A more sophisticated view is, “Did they say yes without external pressure such as alcohol, force, or threat of force?” Does a slightly better job of determining true consent by my definition, but still isn’t quite there.

Better yet, add “implied force, peer pressure, hierarchical pressure(boss, teacher, adult to kid), cultural pressure, or economic pressure”.

However, that can be very hard to test for, and our society’s focus on freedom and being able to go with the flow and not be too analytical can make it hard to determine consent to that level.  Explicitly asking those questions can get you a rejection you might not otherwise have gotten.  Again, this creates wiggle room for people who do know that they wouldn’t have gotten sex without external factors.  The vast majority of rape accusations are against people who knew they were applying outside pressure or that some other factor was.

However, the ethical standard I’m choosing to apply is, “Did the accused (or not, if you’re judging yourself) know that under normal circumstances, the other party would not have consented to sex with them?”.  If so, and if they had sex with the person, they must have known that the person’s capacity to consent was compromised when they decided to pursue sex.  Legal issues aside, this is unethical.  It also often accounts for why people view some approaches to obtaining sex as sketchy or generally less than a stellar recommendation of someone’s character.  If you’re admitting something is sketchy, I’d argue you are admitting there’s a good chance it is either unethical, should be criminal or both.

Now, is that rape?  No, I don’t believe so.  I would restrict rape to the person knowingly applying their own form of force through physical means: i.e., physical force, threat of physical force, implied threat of force, them drugging the person, or them getting the person drunk.  However, I do think it should be considered immoral, unethical, and probably criminal.  The crime here is intentional denial or avoidance of consent for the purposes of obtaining intercourse with the person.  We don’t have a rape problem, we have a consent problem, and insisting on focusing on rape obscures that.  Certainly in our lifetime, it’s unlikely this sort of crime will ever be considered under the umbrella of rape from a legal or pop culture standpoint, and I think trying to shoehorn it into that category makes a difficult task even harder.

Now, onto the world-building section, it is a bit short, since this is an example-based article.  Using this as an example can you think of any other issues that suffer from similar complexity?  There are quite a few.  Drug crimes, religion, various areas of ethics.  The humanities, the sciences.  You can use the contrast between culture, rhetorical value, and linguistic meaning to add depth to any area of your world-building.  The spaces between these related meanings leave people room to rationalize, have different opinions or takes on a subject, and room for cultural change and/or growth.  This also applies to conflict between individual characters and groups of characters.

 

The Translation Problem: People vs. Computers

In my last post, I introduced the topic of natural language processing and discussed the issue of how the context of a piece of language has an enormous impact on its translation into another language.  In this post, I want to address issue with translation.  Specifically, I want to talk how language is really an integrated function of the way the human brain models the world, and why this might make it difficult to create a machine translator isolated from the rest of an artificial intelligence.

When a human uses language they are expressing things that are based upon an integrated model of the universe in which they live.  There is a linguistic model in their brain that divides up their concept of the world into ideas representable by words.  For example, let’s look at the word “pit bull”.  (It’s written with two words, but as a compound word, it functions as a single noun.)  Pit bull is a generic term for a group of terrier dog breeds.  Terriers are dogs.  Dogs are mammals.  Mammals are animals.  This relationship is called a hypernym/hyponym relationship.  All content words(nouns/verbs/adjectives) are part of a hierarchical tree of hypo-/hyper-nym relationships.

So when you talk about a pit bull, you’re invoking the tree to which it belongs, and anything you say about a pit bull will trigger the conversational participants’ knowledge and feelings about not only pit bulls, but all the other members of the tree to which it belongs.  It would be fairly trivial programming-wise, although possibly quite tedious data-entry-wise to create a hypo-/hyper-nym tree for the couple-hundred-thousand or so words that make up the core vocabulary of English.  But to codify the various associations to all those words would be a lot more difficult.  Such a tree would be a step towards creating both a world-model and knowledge-base, aspects of artificial intelligence not explicitly related to the problem of machine translation.  That’s because humans use their whole brain when they use language, and so by default, they use more than just a bare set of grammar rules when parsing language and translating between one language and another.

One use of such a tree and its associations would be to distinguish between homographs or homonyms.  For example, if the computer sees a word it knows is associated with animals, it could work through the hypernym tree to see if “animal” is a hypernym or association with say, the word horse.  Or, if it sees the word “grain”, it could run through the trees of other words to see if they are farming/crop related or wood-related.  Or, perhaps, crossing language boundaries, if a language has one word that covers all senses of “ride”, and the other language distinguishes between riding in a car, or riding a horse, the program could use the trees to search for horse- or car-related words that might let it make a best guess one which verb is appropriate in a given context.

The long and short of the case I intend to make is that a true and accurate translation program cannot be written without taking enormous steps down the path of artificial intelligence.  A purely rule-based system, no matter how many epicycles are added to it, cannot be entirely accurate, because even a human being with native fluency in both languages and extensive knowledge and experience of translating cannot be entirely accurate.  Language is too malleable and allows too many equivalent forms to always allow for a single definitive translation of anything reasonably complex, and this is why it is necessary to make value judgements based on extra-linguistic data, which can only be comprehensively modeled by using techniques beyond pure grammatical rules.

 

In the next post, I’ll talk about statistical methods of machine translation, and hopefully I’ll be following that up with a critique and analysis of the spec fic concept of a universal translator.

 

The Translation Problem

One of the biggest problems in the field of artificial intelligence as it is popularly conceived is that of natural language processing and machine translation.  Natural language processing is the processing of human languages, generally in the form of text, but also as audio, by computers and software.  Machine translation is a form of natural language processing having to do with the translation of a text from one language into another by a computer.  Enormous strides have been made in natural language processing over the last two decades.  Text-to-speech, represented often by those annoying robotic phone calls, and speech-to-text such as dictation software, have both improved markedly.

But the big fish of natural language processing is still the universal translator, able to translate between any two given languages with high accuracy, perhaps even perfectly.  Which is, currently impossible, despite what people might tell you.  For example, Skype Translator looks great in the recent demo, but it’s doubtful it performs quite so well in the field.

There are actually several major issues that machine translation, and translation in general, have to tackle to achieve a high degree of accuracy.  Over the next few posts, I’m going to discuss what some of them are, why they are so difficult to solve, and possible ways they might be overcome.

To start with, there’s a very simple problem with most translation:  language has meaning on its own, but for human communication purposes, you have to know the context of the words you are trying to translate.  Most communication is not factual transmission of information.  Even in formal written prose, there are heaps of subtle ambiguities, and most humans can’t always pick up on every one, much less a poor dumb computer.  This is a problem not only when a word is used that has two different sense, such as a technical and a general sense.  Different languages divide up their semantic space differently, and a word that is a correct translation on paper may fail miserably in the real world.

To give an example, Japanese has pronouns just like English.  However, it does not have an exactly equivalent set.  The word for “her” in Japanese is “kanojo”.  But that word is rarely used, as most Japanese uses a name and an honorific instead.  Further complicating the situation, the word “kanojo” is also idiomatic for “girlfriend”.  The same goes for “kareshi” which is often translated as “he” on paper.   And then we have the various Japanese pronouns technically meaning “you”, or at least referring to the second person singular.  The main one given in teaching Japanese is “anata”.  However, much like “kanojo” or “kareshi” there is a sense of intimacy not immediately apparent in the use of this word.  It is often used between older married couples to mean something like “dear”, and is rarely if ever used as a bare second person, singular pronoun.  Rather, again, the person’s name and an honorific are used.  However, a great deal would have to be hard-coded into a piece of translation software to cover all of this, and that’s assuming there’s even enough context to make a decision at all.  Further, going into Japanese from English, would require the addition of formality information that a computer or even a human couldn’t always be sure of.

It’s often possible to translate conversation/survival language with bare, literal translations, because the basic sense is all that matters.  But more formal registers, or translation of a fixed text, would likely prove impossible for today’s machine translators and most humans, because there’s no way to convey all the necessary context that would provide information that in one language would not be included lexically or grammatically, but it all but necessary for appropriate communication in the other.

 

In the next post, I’ll be looking more at the ways language divide up their semantic space, and how that hinders accurate translation.

 

The Difference Between Spoken and Written Language: Acronym Edition

Something I’ve noticed recently online is the issue of the indefinite articles: “a” vs. “an”.  Many people probably know the rule for this, and many people probably just do it unconsciously.  Essentially, you have “a” before a word beginning with a consonant sound (not consonant letter!), and “an” before a word beginning with a vowel sound (not vowel letter!).  This kind of thing is called “allomorphy”, made up of the Greek roots(morphemes) “allo”, meaning “other”, and “morph”, meaning “shape”(form).  They have different forms depending on the words around them.

Now, there’s an interesting intersection between written and spoken language here:

1. Often, people taught the rule explicitly put “a” before any written word with an initial consonant letter, and “an” before any written word with an initial vowel letter.  There are a few variations of this.  And with dialects, there can be differences, as well.  The “a historical”/”an ‘istorical” debate is still raging, for example.  And then you have the examples like “an apron”, which was originally “a napron”, but because of the ambiguity in speech, people reanalyzed the morpheme boundary to get our modern usage.  The “an (vowel)-” beginning was just so much more common than the “a n–” combination, so people who were hearing the phrase for the first time just assumed one analysis based on their past experience.

2. The issue of whether an acronym should be read as its individual letters, it’s whole word pronunciation, or the entire phrase that it represents.  For example, should the indefinite article for the new age category in publishing, “New Adult”–acronym “NA”–be written “an NA” or “a NA”.  The first version would be correct if it was being read “en ey”, but the second would be correct if it was read “New Adult”–despite being written in acronym form–and although it doesn’t apply for this case, if “NA” was a true acronym instead of an initialism, you could argue it should be “a nah”.

 

Personally, I would never read “NA” as “New Adult” out loud, and so seeing “a NA novel” confuses the heck out of me for a second or two.  But other people seem to think that’s a legitimate reading, and who am I o gainsay them?  I wonder how this might apply in an editing situation, where the editor and the writer disagree about which is the proper way to read an acronym.  Or in a critique?

 

Linguistics and SFF: Appropriation and Dialect

Last time on Linguistics and SFF: Orthography and Etymology

An oft-debated topic in all fiction is the subject of using dialect as dialogue.  Many famous writers have done it, and many not-so-famous writers have tried it, to varying degrees of success.  Since dialects are a very linguisticky topic, I thought I’d take a look at why and how writers use them, some of the effect of using them, and how it all relates to the whole debate on cultural appropriation.

First, a few thoughts on dialect:

Definitions

1. A dialect is a unique language system characteristic of a group of speakers.

2. A dialect is a variety of a major language carrying connotations of social, cultural, or economic subordination to the culture which speaks the dominant language.

These two definitions exist simultaneously.  For our purposes, the second one is the most relevant.

Dialects under the second definition are culturally, socially, and economically stigmatized by the dominant culture.  Speaking a dialect is often portrayed by dominant cultural institutions as just “bad [Dominant Language]”.  For example, “bad English”. (We’ll keep this example, since I’m discussing literature from primarily English-speaking countries.)  Many children are taught in school to speak “proper English” in school, and punished for using their native dialect.  “Proper English” actually takes a few forms:  In America, it is a dialect known as “Standard American English(SAE)”, which is most similar to Midwestern dialects of American English.  In England, there is Received Pronunciation(RP).

Most other countries with official “languages”, have a similar pattern of official and unofficial dialects.  What is considered a language is often up to whichever dialects can get state support, and it has been said that “A language is merely a dialect with an army and a navy.”  Or, in the case of France, “a dialect with a national Academy.”

Almost all other dialects are usually considered inferior or degraded versions of the official dialect.

So, onto the use of dialect in fiction.

For the most part, dialogue in English-language novels is written in the standard form of written English, which reflects more or less the standard form of spoken English in the country in which it is printed.  Although, depending on the orthography used, this reflection could be rather cloudy or warped.  Dialect, then, is represented in an attempt at “phonetic” spelling and non-standard vocabulary and grammar.

Most commonly, because the author does not often speak the dialect natively that they are attempting to represent, dialect in fiction falls back on stereotypes of usage related to the cultural perception of the spoken dialect.  This can lead to a continuation of prejudice and stereotypes, and is also a form of linguistic and cultural appropriation, as a member of the dominant culture makes use of minority culture for their own ends.  Rarely in the cases we’re examining are these ends malicious.  But they are often still quite problematic.

There are many English dialects that have been popularized in mass culture, with varying degrees of difference presented.  For example: Italian American English, Chinese American English, African American English, Cockney English, Appalachian English, and Southern English.  In fact, they are so parodied, mocked, and appropriated that they have “accents” associated with them.  The cheesy Italian accent a la Mario, the “Oy Guvnah” of Cockney, and “tree dorra” of Chinese American English.

Some of these “dialects” are actually accents or inter-languages, rather than stable dialects.  However, they are all commonly referred to as “dialect” (or occasionally “accent”) in regards to their representation in fiction.  And for the most part, rather than actual depictions of the stated dialect, what is really present is the set of stereotypical markers associated with the dialect by mainstream culture.

Next time, I’ll look at some examples, both made-up and used in novels, of dialect appropriation.

Next time on Linguistics and SFF: Artemis Fowl and the Eternity Code

 

Linguistics and SFF: Orthography and Etymology

Last time on Linguistics and SFF: Orthography and Vowel Systems

In this post, I’ll be discussing orthography and etymology.  Etymology is the study of the history of words–basically tracing the forms of a word through time all the way back to its origins.  Since I’m writing in English, I’ll mostly be using English examples for etymology.  And that’s great, because English has such a complex past.

Now, where the orthography part comes in is that you can often tell the origin of a root in English by the orthography.  English spelling will use a letter combination for words from one language to represent a sound, even though we already have a perfectly good letter for that sound.  And its based partially on how orthographies interact, and partially on language change.

So, English has several alternate spellings based on etymology.  For example, the use of “ch” instead of “k” for the /k/ sound.  An example is the word “chemical”, which comes from Medieval Latin “alchimicus” through the word “chemic” in the late 1500s.  It goes back to the Greek form “khemeioa” through Arabic “al-kimiya”.

“Ch” is called a “digraph”, meaning two graphemes used to represent a single phoneme.  It is common to many roots received originally from the Greek.  These roots often passed through Latin,which used the “ch” digraph to represent the Greek /kh/ an aspirated voiceless velar stop, which Latin lacked.  The pronunciation became /k/ in Late Latin, and is thus the one we’ve inherited into English today.

I go into detail here for a few reasons.  First, understanding your own orthography can be useful ind designing others.  Second, most novels use strictly English orthography, so options are limited.  However, by considering the connotations of various loan word roots in English, you can achieve a certain amount of meaning.  For example, in order to differentiate a con-word in a novel, you can use common digraphs such as “ch”, even if the word would be pronounced with a normal /k/ sound.  In this way, you can make a word seem older, foreign, etc without resorting to special symbols, or that mainstay of conlang/foreign language typography, the italic word/phrase.

You can use similar digraphs/letter choices to create other differences.  Although it’s arguable that this is an appropriation rather than a proper usage, Jay Kristoff, in his book Stromdancer, spells the Japanese word “salary-man” “sarariman” in imitation of the most common Japanese romaji transliteration system, in order to make it seem more foreign.  I’d argue, however, that the clear English root makes a good case for spelling the word in English, to make clear its loan word status.

And this leads us to the concept of transliteration, which is spelling a word in Roman letters that normally would not be so spelled.  Chinese and Japanese both have a multitude of English transliteration schemes, such that the same word can look vastly different and be pronounced completely differently by a native English speaker.

This fact can be used quite effectively to manipulate orthography to create certain effects.  Really the only limit is your imagination.

Next time, we’ll be taking a bit of a side-trip while I talk about the use of “dialect” in speculative fiction–it probably applies to any type of fiction, though.

Next time on Linguistics and SFF: Appropriation and Dialect

 

Linguistics and SFF: Orthography and Vowel Systems

Last time on Linguistics and SFF: Orthography and Exoticism

Last time, I discussed some basics on orthography and its function in writing systems.  I’ll be doing a bit more of that in this post, but in the context of using that insight for the purposes of writing speculative fiction.  I’ll be focusing, as the title suggests, on what the vowel system in a language suggests about the best orthography for that language.

First, considering the structure of the three major types of writing system, they have advantages and disadvantages, and these are in part based on the structure of the language to be represented.

The difference between alphabets and syllbaries is much smaller than the difference between them and a logography.   They both represent the sounds of a language rather than its meaning.  So I’ll be focusing first on them, and why you might want to us one over the other.

An alphabet is best for a language with a large number of sounds(phonemes).  English, for example, has a great many sounds, and not only that, it has far more vowel sounds than most languages.  English is most commonly described as having about 16 distinct vowel sounds.  The most common system of vowel sounds in the world is a system of 5: /a/ /e/ /i/ /o/ /u/.  English has equivalents of all those vowels.  I bet you can guess what they are.

The number of vowels matters for a very specific reason.  Vowels differ from consonants in that a consonant phone or phoneme has a very limited variety in its articulation.  Consonants have four major features:

1. Place of articulation describes where in the mouth you place your tongue when making the sound.  These are generally divided into laryngeal, the back of the throat, velar, the soft palate, palatal, the hard palate  alveolar, the ridge behind the teeth, dental, the teeth, and labial, the lips.

2. Manner of articulation describes how the breath is released during speech: stop/plosive, when the tongue/lips create a closure or “stop” of air, and the air is then released, affricate, where there is a stop, followed by a protracted release, fricative, where there is no closure, but rather a proximity of the tongue/lips to the place of articulation and a protracted release of air, and sonorant, which (basically) is everything else.

3. Voicing describes whether the vocal folds(vocal chords) are vibrating during sound production.  The major voicings are voiced, a clear vibration, unvoiced/voiceless, no vibration, and some other, less common voicings.

4. Aspiration describes the strength of the airflow.  It’s generally divided into aspirated, with a noticeable puff of air, and unaspirated, with no noticeable puff.

Now, these features are in general fairly binary, either “on” or “off”.  And that means you can’t vary your phones too much without changing your phonemes.  But vowels are different.  Vowels have basically two features.  However, there is quite a bit of room for variance, and so a sound that one speaker would perceive as an English short “e” might be considered by another to be an English short “i’.

The point being, that most languages have smaller systems of vowels because they divide the space up less finely.  It is possible to have a language with as few as two vowels, and three is not uncommon.  There are also many systems with seven vowels.  So, English is rather uncommon in the number of vowels it has.

You might have figured out already why the number of vowels is significant: even English with its many vowels and large number of consonants has only about 50 phonemes at most.  Fifty letters is not an over-large number to memorize.  But what about syllables?  With 12 vowels, 25 consonants and just using a (C)V syllable structure there are about 300 possible syllables. And English has fairly permissive structural rules for syllables, so the real number is much greater.  By contrast, Japanese, a language which uses a syllabary, has that five vowel system we mentioned before, and about 12 consonants.  That’s a little over 60 possible syllables with its (C)V(n) structure.  And even then, Japanese has shortcuts.  Consonant voicing uses a diacritic mark rather than a whole second set of syllables, and rather than having syllables with the coda (n), uses a single symbol for it, which is simply added after the (C)V syllable variant.  So in order to have a syllabary, a language must have relatively simple phonotactics, few phonemes, especially vowel phones, since vowel and consonant numbers are multiplicative under a syllabary.  Hawaiin, for example, has five vowels and 12 consonants and a (C)V syllable structure, and is thus perfect for a syllabary, whereas German, with its vowels and complicated consonant clusters is not.

A logography, of course, has no such issues, since it doesn’t care about sounds.  Logographies are great for languages with little morphology, since they aren’t alterable for tense or number, but require separate symbols for each.  They do, of course, have the same issues with massive memorization that an English syllabary would have.

Since writing systems tend to develop based on the languages they represent, you can do some fun and efficient world-building based on how you choose to represent a language orthographically.  Obviously you can’t really do much about an actual writing system in a novel, but it is something to keep in mind, especially when using a real-world language that doesn’t actually use English orthography.  When you think about that language and how the various characters feel about it, you can use that in description.

For example, many people think Chinese orthography is particularly beautiful or elegant, and the discipline of Chinese calligraphy arose from the type of writing system Chinese uses.  Whereas though English has forms of calligraphy and also medieval illumination, it is not seen in quite the same light.

In the next post, I’ll be discussing two things: etymology and orthography’s effect on the reality and perception of it, and also how you can use your chosen orthography to illuminate language and culture and also how when writing a real-world language you aren’t a native speaker of, you may mis-understand or mis-represent it because of your own prejudices and pre-conceptions based on your own language having a different writing system.

Next time on Linguistics and SFF: Orthography and Etymology

 

Linguistics and SFF: Orthgraphy and Exoticism, Part 1

Last time on Linguistics and SFF: Language and Exoticism

In part one of this post, I will be giving a short introduction to several linguistic concepts relating to writing systems.  This is somewhat technical, but that’s kind of the point of the post series.  The actual application of these concepts to speculative fiction will take place in the second post.

[] shows a phonetic representation

// shows a phonemic representation

<> is the identifier for a gramphemic/orthographic representation, but I’ll be using double quotes for simplicity.

The technical term for the way a language is presented visually is a “writing system.”  Writing systems are made up of pieces called “graphemes”.  Each grapheme represents a discrete element of the written language–generally they each represent one of the smallest semantically distinguishable unit of the written language.  Together, the set of graphemes used to visually represent a language make up that language’s orthography.  Technically speaking, an orthography is a standard system of graphemes, but there are plenty of non-standard systems.

Graphemes, and therefore orthographies tend to come in three major types:

1. An alphabet is an orthography where each grapheme (at least theoretically) represents a single sound, or phoneme, in the language.

2. A syllabary is an orthography where each grapheme represents a syllable in the language.

3. A logography is an orthography where each grapheme represents a word or morpheme in the language.  It’s also a word I just made up–although it might exist somewhere–through derivational morphology.

Now, since a lot of people have trouble with this and many lay people use a non-technical form of linguistic terminology, I’m going to explain what phonemes, syllables, and logograms are.

A phoneme is the set of phones–individual sounds formed by articulation of the human mouth–that a language considers to be the same “sound” for the purposes of distinguishing words.  The two major types of phoneme are the consonant and the vowel.  They are demonstrable by a minimal pair, a pair of words considered to have different meaning which can be used to show the phonemic difference between two phones.  For example, “mitt” and “kit”, with the phonetic representation of [mit] and [kit].  The difference in meaning is distinguished by the two phones at the beginning of the words.

The symbol for a phoneme is called a letter.

A syllable is a collection of sequential sounds made up of a vowel and the consonants associated with it.  Every language on earth has a set of rules–called phonotactics–which describes the possible structures of syllables in a language.  Japanese, for example, has a syllable structure consisting of (C)V(N), where C represents any consonant, V represents any vowel, N represents a nasal consonant–in the case of Japanese, these are /n/ and /m/–and parentheses mean the phoneme is optional in that position.  Thusly, “kan” is a legal Japanese syllable of the structure CVN, and kar is an illegal syllable, because the structure is CV!N, where the “!” means the following consonant type is not accurate.

A few technical notes on syllable structure: The group of consonants before the vowel is called the onset.  The vowel is called the nucleus.  The group of consonants following the vowel is called the coda.  The combination of the nucleus and the coda is called the rhyme–hopefully obvious reasons.  You’ll understand why I mention this in the analysis post.

The symbol for a syllable is called a syllogram for the purposes of this series.

A logogram is a grapheme representing a word or morpheme in the language.

Now, these types of system are theoretical constructs.  Very few systems are exactly as described above.  For example, English has an alphabet, but there is not a 1-to-1 correspondence of letter to phoneme.  So, despite what you may hear, English is not a phonetic or even phonemic writing system as the various IPA(International Phonetic Alphabet) alphabets are.  And Japanese makes use of both syllograms and logograms.  And even then, some words in Japanese that are written in syllograms(kana) use a seemingly incorrect syllogram.  For example, the topic particle “wa” is spelled with the syllogram for “ha”.

English is an example of a type of non-phonetic(phonemic) alphabet called a morphophonemic alphabet.  For example, the English words “electric” and “electricity” are spelled with the same letters, even though phonemically, they should be represented as “elektrik” and “elektrisity”  The same is similar for many morphemes with Greek or Latin roots where derivational inflection–the creation of a new word by the systematic addition of morphemes–is applied.  This is because advanced reading is not done by sounding out each letter, but rather by recognizing the shape of the word.  This is why the two gimmicks or removing the vowels in a word, or scrambling a word while keeping on the first and last letters the same don’t interfere significantly with the readability of the text.  This is a valuable tool in a writing system, even though it may make learning that system more difficult.

Another example of the oddity of English spelling is the famous “alternate spelling” of fish–“ghoti”, often attributed to George Bernard Shaw.  It was used to highlight perceived “irregularities” in English spelling.  However, it is based on mis-understandings of how spelling works, in English and in general.

“gh” only makes the /f/ sound in the coda of a word.  “enough”.  It still makes a /g/ sound in the onset position: e.g. “ghost”.  In the so-called word “ghoti”, the “gh” is in the onset position, where it makes a  “hard g” like in ghost.

The “ti” is supposed to represent the “sh” sound.  However, it does so only in the “medial” position–that is, between two other syllables.  It always makes the “ty” sound in the final position as in the word “ghoti”.

Finally, the “o” is supposed to make the “i” sound, as in “women”.  However, that is not a regular pattern of English, but the fault of some other very simple rules.  In an unstressed syllable, such as the “-men” and “-man” of “women: and “woman”, the sound represented by the letter “e” becomes a schwa instead of an /e/ or /a/ as in the stressed syllables of the male forms “men” and “man”.  This causes the words to be indistinguishable in spoken English.  So, because the rule of a vowel becoming a schwa in an unstressed syllable has priority, the “o” in the stressed syllable is altered to distinguish between the singular and plural forms.

But as we learned above, English writing is a morphophonemic alphabet, meaning that it works to preserve the clear relation between two allomorph–different forms of the same morpheme–over maintaining phonemic accuracy.

Further, writing systems in general are made static at a certain time in the languages development.  So, because writing systems care less about the accurate phonemic representation of a word than of readability, when the spoken language evolves, the written language does not, creating these odd orthographical artifacts.  And this is visible in many writing systems, not just English.  Something to consider in the follow-up post, and while deciding on the orthographic representation you choose to use in your story.

Next time on Linguistics and SFF: Orthography and Vowel Systems

 

Linguistics and SFF: Signs and Signals and Story-telling

In my past few posts, I have been using the words “signal”, “sign”, and “signifier”, and rather haphazardly at that.  But these are real linguistic concepts, and that have concrete meanings.  And, perhaps more relevant to this post, these concepts can be useful tools in constructing not only your speculative world, but the character arcs and themes underlying your narrative.

So, first some definitions:

Language is made up of a set of signs and the rules that govern their interactions.  You can, for most purposes, consider a sign to be a word in context.

Every sign has three(1) parts:

The signifier is the group of sounds (or letters, for written words) that make up the word.

The signified is what we think of when we hear or see the word.

The referent is what the word refers to in the context in which we encounter it.

For example:  we have a sign with the signifier “president”.  It’s signified is the head of the executive branch of the US government. (Strictly speaking, that’s just its denotation.  You can include the connotation, as well.)  And the referent currently would be Barack Obama.  In 1992, the signifier and signified would have been the same, but the word would have a different referent when used in the US in the absence of any modifiers.

Now, when working in speculative fiction, you can use these aspects of a word (or sign) to decide what English word to use, or if you might like to use a foreign word.

For example, what do you want to call your political units?  We have many choices for this, many possible signs.  We have empires.  An empire is a political unit consisting of many smaller states brought together by forces–most commonly conquest.  An empire can have connotations of bureaucracy, soaring capitals and primitive backwaters.  There are many real-world referents, such as sophisticated Rome, far-flung Alexandria, brutal Azteca, and powerful Britain.

And you can capitalize on aspects of those various referents, much as you could capitalize on the various connotations.  Or, you could pick a kingdom, and capitalize on the myriad kingdoms throughout history.  The same is true for anything, from religion, to clothing, to character.

In the next post, I’ll look at how you can make use of the written form of a sign in your world-building, and I’ll give some examples of foreign signs and how these concepts can influence their use.