Mega Post Thanksgiving

(The first part of this was supposed to be posted on Monday, but the server crashed and deleted half of it. Hence the extra long post.)

[commence tryptophan-less rant]

Over break, I came across two interesting language related articles, both of which dealt with birds.

The first was a short summary of research done by researchers at the University of Chicago and UCSD, where they taught European starlings “two syntactic patterns, one thought to be understood only by humans.” Presented with novel patterns, the birds could identify which ones followed which grammatical structures. The excerpt isn’t online, but the original research article can be found here. The researchers used two different sounds arranged in two different syntactic patterns: A^nB^n and (AB)^n, the latter presumed to be understood solely by humans. The paper ultimately concludes that the songbirds can, in fact, differentiate between the two patterns, but from reading the article, a great deal stands out. For example, the birds only began to differentiate after operant conditioning had taken place, and that too after 10,000-50,000 trials. The article points out that tamarin monkeys, given a similar task, failed to recognize differences between the two patterns, but given how essential patterning is for songbirds, the result doesn’t seem terribly surprising. Moreover, this only points the ability to identify recursive grammar, not actually use it. It’s akin to me being able to identify an airplane, though I have no idea how to actually fly one. If research showing animals actually using recursive grammar surfaces, then there would be something interesting to talk about, but at present, this article doesn’t offer much.

The second article I happened upon over Thanksgiving break was much more seasonal. It was in The New Yorker, so I can’t find it online, but it was about a man who claims he can talk to turkeys. Whereas experts claim turkeys have 13 basic sounds, Joe Hutto of Florida says he’s mastered 50 distinct sounds. Over years of rearing toms and hens from eggs, he’s honed the ability to talk, console, and instruct turkeys, as well as listening and learning from them. Of course, there’s no empirical evidence for this but it brings up the possibility of animals having some basic communications system. So while birds of a feather fly together, Hutto may not be so cuckoo after all.

[end tryptophan-less rant]

More recently, I read an artice on how boys and girls process language differently (once again gender difference is brought up). The article describes this study, where it was shown that girls over-generalized more often than boys. Over-generalization, in this case, means (for example) saying holded for held. The research examined such patterns among children of various ages and found that the over-generalization rate was three times higher among girls than in boys. This ran contrary to the hypothesis, since females generally outperform males on verbal skills. The proposed explanation is that girls created past tense conjugations based on similar-sounding words (e.g. folded and molded), whereas boys derived such formations based on grammatical rules and exceptions. While this explanation may suffice, it does not explain why these differences should exist in the first place. The study itself is interesting (albeit abstruse) and opens the door to other interesting research. Another note about this article is that it focused on a specific, measureable aspect of language, as opposed to vague generalizations, in discussing gender differences. It’d be nice to have more similarly-designed studies.

Annoying Vowels

(I had another post drafted, waiting in the wings, but my increasing frustration with IHUM led me to publish this one.)

In Visions of Mortality, we’re currently reading the Bhagavad-Gita, a tiny (yet pithy) excerpt of the Mahabharata, a colossal Sanskrit epic of familial strife and redemption. It’s an engaging, interesting, and enlightening tale, but longer than the Odyssey and Iliad combined. I had the fortune of being exposed to the televised version as a kid. True, it was over 72 hours long, but it’s probably faster than reading. In any case, I was exposed to it in Hindi, so all my knowledge of pronunciation was based off native pronunciation schemes.

Fast forward nine-ish years. In my hands lie the English translation of the Bhagavad-Gita and to my horror, I find most of the names are appended with a rather annoying -a. From what I’ve been taught, however, the final -a is rarely pronounced. So while most Americans pronounce “Arjuna” in three syllables, virtually all Indians would sound it as “Arjun.” It’s a pattern prominent in all male names, so today I thought I would look into this puzzling problem.

One thing to keep in mind is that the epic was originally written in Sanskrit. As I discovered, Sanksrit has a general convention whereby when a noun ends in a consonant, a schwa (short neutral vowel) follows it. There are exceptions, most notably when the original word ends in a long vowel. However, as Hindi slowly evolved from Sanskrit, it dropped the schwa entirely from Sanskrit loan words. This is the pronunciation I learned, and it’s the way an entire subcontinent pronounces it. If that’s the case, then why does this problem exist in the first place?

Well, as early as 1816, there has been interest in Romanizing and transliteration Sanskrit from the Devanagari script. As you can imagine, the vowels and consonants didn’t exactly match up, so several different schemes have been proposed. The current de facto standard is from the International Alphabet of Sanskrit Transliteration (IAST), which uses special characters to provide lossless transliteration of sounds (sample: yogaścittavṛttinirodhaḥ), making it difficult to learn and type. Two other popular transliteration methods are Indian languages TRANSliteration (ITRANS) and Harvard-Kyoto (HK). Both of these, however, use capital letters, which can create confusion when writing names (ITRANS sample: yogashchittavR^ittinirodhaH; HK sample: yogazcittavRttinirodhaH). Both of these have the upshot of being keyboard friendly though.

People crave ease of use, however, so approximations using the Latin alphabet and English pronunciation have been attempted (sample: yogash chitta-vritti-nirodhah). This has, in my experience, led to two problems: first, highly inconsistent transliteration patterns; and two, the now-redundant schwa. With Latin transliteration, this short vowel is symbolized by an -a appended to a noun. For the Western reader, however, this gives the false impression of a short a sound following every masculine noun, for both Sanskrit and Sanskrit loan words in Hindi. Hinduism has a long tradition, so many people and place names have roots in Sanskrit and now are Romanized with an extra -a. The end result? Frustration and anguish for native speakers like myself.

This brings up a larger consideration: Sanskrit is a dead language, so should it be Latinized from it’s original pronunciation or from it’s de facto standard Hindi pronunciation? Before you jump on the notion of preservation-for-preservation’s sake, I’d like to remind you of how many Latin names have been Anglicized. After all, who among us actually pronounces Caesar as /ky-sar/ and Cicero as /kee-ker-o/?

Languages change, pronunciations shift, and humans keep up with it. Unfortunately, the overwhelming tradition of transliteration has placed the -a in a position where I can’t see it ever disappearing. So it falls on us to make sure we’re pronouncing words, in any language, correctly. At the very least it’s a sign of education, and at the most it’s pure respect.

Further reading:
Wikipedia article on transliterating Sanskrit
Detail of difficulties with Hindi pronunciation (scroll down to right before the chart)
Comparing different systems Sanskrit of transliteration
The Encyclopedia of Authentic Hinduism, which employs a revised and more accurate system of transliteration

Language Tidbits

I came across two really interesting articles today. I shall briefly describe them, for the sake of time (I nee to be awake in sub-9 hours).

1. ISN: A very interesting story about Nicaraguan deaf children who, having never learned any formal language, when brought together naturally created Nicaraguan Sign Language. It parallels much of what was talked in class, plus it sheds light on language acquisition theories.

2. Liet-Lavlut: Recently held for the third year in Östersund, Sweden, it is the only international music contest for minority languages. Among the languages represented at this year’s event include Furlanic, Occitan, Votic, and Meänkieli. If this gets greater international attention, perhaps language revival will really take off.