RSS Feed

Through the Language Glass: A Conlanger’s Perspective

Posted on Tuesday, October 19, 2010 in Books, Conlanging, Nonfiction, Rant

Through the Language Glass: Why the World Looks Different in Other Languages, the most recent book by Guy Deutscher, is a thought-provoking, highly-readable romp through some of the more unexpected realms of language. Deutscher is more than happy to take on with relish those ideas considered orthodox or “common sense” and to defend his own views, backing them up with facts and research. A number of very positive reviews of Through the Language Glass can be found online (such as here, here, and here, just to note a few), but what does the work have to offer conlangers? Being primarily an artlanger, I apologize ahead of time if this review strays too far in that direction. Deutscher’s book examines three main areas in-depth: color terms, grammatical gender, and spatial orientation. Let’s look at each of these in turn.

First, we examine color terminology. Deutscher examines how different languages split up the spectrum, and it turns out that my *blue* may not necessarily be your *blue*. In explaining the history of color terms, the author reaches back to W. E. Gladstone’s Studies on Homer and the Homeric Age, Volume 3, and its section entitled Homer’s Perception and Use of Colour and continues all the way up to the World Color Survey. The idea of slicing up the WCS-Munsell Color Chart or the CIE Color Space in novel ways to highlight the alien-ness or foreign-ness of a conlang’s speakers appeals to me. As Deutscher points out clearly, there are a number of interesting variations on the theme of how colors are designated in the world of natural languages. Even a language as familiar as Russian has "синий" and "голубой" whereas English simply has "blue". Granted, we English-speakers have shades of blue (light, dark, navy, baby, sky, etc.) and blue-related colors (turquoise, azure, indigo, etc.). Is purple a kind of blue? These are all concepts that can be played with in a conlang.

Next, let’s look at grammatical gender. Personally, I think this is a largely untapped field for conlangs to play in. Some of the African and Australian gender systems are very tempting to riff on. Deutscher talks about the “Bantu languages such as Swahili [which] have up to ten genders, and the Australian language Ngan’gityemerri is said to have fifteen different genders.” Of the 319 languages which specify use of grammatical gender in the Conlang Atlas of Language Structure database, over half (175 or 54.8%) use no gender at all. More interestingly, 84 conlangs use the familiar sex-based gender system and only 60 are non-sex-based. Another thing to consider is that it might be fun to determine how gender is assigned to words in one’s lexicon. One of my favorite stories related by Deutscher (in both Through the Language Glass as well as The Unfolding of Language) is how the Gurr-goni word "erriplen" (English "airplane") ended up in the vegetable gender category. It makes perfect sense when the history is known (or surmised), but, on first glance it makes no sense at all. Of course, if one is attempting to streamline or simplify language, gender is probably first on the list of things to excise. Alternatively, a conlang that attempts to be akin to natural languages would be more than likely to have grammatical gender. Why not have fun with it?

Finally, what about spatial orientation? This is the area that made the most impact on me. I really want to come up with a novel (or one that at least doesn’t have too many ANADEWisms) orientation system now. Deutscher provides several excellent alternatives to simple left-right-front-back. One is Guugu Yimithirr which uses an east-west-north-south system all the time. Another is Tzeltal (a Mayan language) which uses a downhill-uphill-across (in the direction of X). On the Marquesas Islands in French Polynesia, one finds mountainward-seaward-east-west. Depending on the geographical location of one’s conlang-speakers, I can see these or other systems being a lot of fun to create. Is there a specific location of historical/spiritual/geographical significance to your speakers? Orient towards that. This would also have some interesting connotations within a story as well (as well as lots of opportunity for misunderstandings… which could be a positive plot device).

These are just a few of the highlights in Deutscher’s new book. His The Unfolding of Language is one of my favorite narrative nonfiction books on language. (Arika Okrent‘s is, of course, another.) His new one is well worth the read, and has some great quotes, intriguing points, and enjoyable prose. Whether you buy it, check it out from your local library, or borrow it from a friend, make sure to put it on your conlang reading list.

Bring on the comments

  1. Keri says:

    Thanks for this great review! I own “The Unfolding of Language” and LOVE it, but I wasn’t sure whether or not his new book would really be of any use to me. Now I want to get it =) Is your link an affiliate link? I’ll be sure to click thru here when I get it!

  2. The Conlanging Librarian says:

    I’m so glad you found the review helpful! I don’t think you’ll be disappointed in the book. The links to Amazon are affiliate links of the Language Creation Society. Happy conlanging!

  3. Jim Henry says:

    “Alternatively, a conlang that attempts to be akin to natural languages would be more than likely to have grammatical gender.”

    According to WALS feature 31, only about 44% of the languages in their sample have any kind of gender system; and of those, about 75% have sex-based gender. I’m not sure offhand how WALS sample languages from different families; if they have one language from each sub-family of large and well-documented families like Indo-European, that could heavily bias the sample toward sex-based gender (and toward gender in general) compared to if they have one language per family, period.

    Anyway, I agree that unusual gender systems are a cool thing to see in a naturalistic artlang. David Peterson’s Zhyler has a good one. Our non-naturalistic collaborative artlang Naeso has IRC-based pronominal gender, with distinctions based on whether an addressed or mentioned person is new to a chat channel or has been participating for a while, whether they’re a regular or infrequent visitor or never online; but so far there’s no gender on nouns.

  4. I figure that conlang gender systems should be based on what ever the culturally important things are. Digital things, analog things and things that taste good.

    I love the idea of grammar as a byproduct giving someone an internal compass. If the language required estimating all quanitities, people would get really good at estimating. For example, crowd covers 4-infinite people, if there were say 20 words for a crowd for different ranges, people who used such a system would be able to estimate the number of pennies in a jar better than mere English speakers. So I think.

Leave a Reply

*