Don't Sleep, There are Snakes

by Everett, Daniel L.

  This distinction between some sounds according to their position in the syllable is important for linguists but is not meaningful in English—any English speaker will understand pin and spin whether or not the p’s in the words are aspirated.

  In sheet versus shit, the difference in sound for a native speaker of English is found in the tenseness of the tongue in producing the vowels. However, since the second word uses a vowel not found in Romance languages, like Spanish or Portuguese, speakers from these languages can have embarrassing moments trying to distinguish such words, since the two sounds are separate phonemes in English but not in Spanish or Portuguese.

  For an inventory the size of Pirahã’s, words need not be quite so long as we might expect, however, because of two additional tools: context and the tones I mentioned earlier.

  Context helps to distinguish meanings in all languages. Consider the English homonyms to versus two. If I ask you, “How many did you say?” and you respond tu (spelled phonetically), we know that only two and not to could be meant in this context. In fact most potential ambiguities are avoided by context.

  I sat one day with Kóxoí at the table under my thatched roof in order to learn more about the sound structure of Pirahã words. Keren appeared with a cup of coffee. She gestured at Kóxoí to see if he would also like a cup. Kóxoí smiled and said, “Tí píai,” which I immediately guessed to mean “Me too.”

  To check this out, I organized a few elicitation sentences to confirm my hunch, acting out and saying, “Kóxoí drinks coffee, Dan píai,” “Kóxoí drinks coffee, me píai,” and so on.

  I recorded examples and isolated the phrases for me too, you too, her too, and so on. Then I asked Kóxoí to repeat them to me so that I could verify their pronunciation.

  What he gave me was surprising and confusing.

  He repeated, “Tí píai.”

  I repeated.

  He said, “Right, kí píai.”

  “What did you say?” I asked with frustration and surprise. Why was he changing the pronunciation? Was there a more simple expression than I had thought?

  “Kí kíai,” he repeated.

  Now I was beginning to question my own sanity. Three different pronunciations in three repetitions. I was sure that the k sound, the t sound, and the p sound were meaningful units of speech—phonemes—of Pirahã. Phonemes are not supposed to be interchangeable! Change Tim to Kim to Pim in English, for example, and you don’t just get alternative pronunciations, you get separate words.

  “Kí kíai?” I asked.

  “That’s right, pí píai” came the exasperating answer.

  In other repetitions, Kóxoí then gave additional pronuciations (again, the x represents the glottal stop of Pirahã): “xí píai,” “xí xíai.”

  I wondered whether Kóxoí’s pronunciation was just “sloppy” compared to other Pirahãs or whether some deeper principle of the language was being illustrated by this variation. It might be that this word was changing meaning in ways that I wasn’t perceiving. Or it could just have been an example of “free variation”—a nonmeaningful difference in pronunciation, as inmy Southern Californian alternative pronunciations of economics as “eeconomics” and “ehconomics,” with no intended differences in meaning. I eventually concluded that it was indeed free variation.

  I observed several other examples of this variation from many speakers. Some people gave me many pronunciations for a single word, like xapapaí, kapapaí, papapaí, xaxaxaí, and kakakaí for the English word head. Or xísiihoái, kísiihoái, písiihoái, píhiihoái, and kíhiihoái for liquid fuel (kerosene, gasoline, butane, etc.).

  I saw that Pirahã allows an astonishing range of variation among consonants. This surprised me, especially in a language with so few phonemes. But at the same time I discovered that Pirahã makes such extensive use of tone, accent, and the weight of its syllables that the language can be whistled, hummed, yelled, or sung.

  For example, the sentence Káixihí xaoxaagá, gáihí (There is a paca there) has a musical form. It is this musical form that is whistled or hummed or sung.

  The vertical lines in the example indicate word boundaries. The notes between the lines are the musical representation of one word. The caret (^) under a note indicates that the syllable is louder than other syllables in the same word. A whole note (hollow oval) represents the longest syllable type in the language (consonant + vowel + vowel) and a quarter note (black oval) represents the shortest syllable in the language (consonant + vowel). The other notes and length dots represent syllables of other lengths—Pirahã has five syllable lengths. The relative height of the notes, that is, the syllables, indicates the tone. A higher note is a syllable with high tone. A lower note is a syllable with low tone. A connecting line between two notes, a musical “tie,” indicates that there is a movement from a low tone to a high tone or a high tone to a low tone, with no pause between the two end points. In the musical representation of káixihí, the first group of notes has a falling tone, followed by a short low tone, with a preceding break in the whistle (where the glottal stop would have been in káixihí), followed by another short break (where the h would be) and a short high tone, and so on. The word, without consonants or vowels, is stressed (volume-adjusted) according to syllable weight. Thus, the syllable boundaries are clearly present in whistle, humming, and yelling channels, all of which figure in this musical representation even though the phonemes themselves are missing.

  There is no staff in my music analogy because Pirahã tones do not have precise pitches like musical notes (middle C on a piano keyboard produces sound waves of 256 hertz), but are relative. A high tone in Pirahã or any other tone language is not a specific number of hertz, but is simply made with greater frequency of vibrations in the vocal cords than a low tone.

  I began to feel that there was a connection between the small number of phonemes and the presence of these “channels of discourse.” I hypothesized that these channels were the key to understanding both the small number of consonants and vowels in Pirahã and the astonishing variation among the consonants. Since all these channels rely crucially on the fact that Pirahã words can be represented musically, we should try to understand a bit of what that musicality derives from.

  First there are the tones. Each vowel in every word has either a high tone or a low tone, in a way similar to Chinese and other tone languages.

  Linguistic tones evolve from a ubiquitous property of the world’s languages, pitch, the relative frequency of vibration of the vocal cords. All languages use pitch to distinguish meanings. In English, for example, a rising pitch at the end of a sentence generally indicates a question, whereas falling pitch indicates a statement.

  In English punctuation, a period is used to mark falling pitch, while the question mark indicates rising pitch. When pitch is used to distinguish the meanings of sentences in this way, it is called intonation. There are many variations of intonation possible. To see just a glimpse of complexity in English’s use of pitch and stress, consider one of my favorite examples, known among linguists as “stress clash override.” When uttered alone, the word thirteen has higher pitch on the last syllable—“thirTEEN.” And the word women has higher pitch on the first syllable, “Women.” But put the two words together and what do you get? You don’t get “thirTEEN Women,” you get “THIRteen Women.” Why?Because English, like several other languages, doesn’t like two high-pitched, or accented, syllables next to each other. It prefers to have an alternating pattern—ACCENTED nonaccented ACCENTED nonaccented, and so on. So speakers of English change the accent on words like thirteen when they precede and modify other words, in order to get the result of alternating accents while at the same time maintaining the accent on the main word of the phrase, in this case the noun women in the noun phrase THIRteen Women. And no English-speaking child ever had to be taught this pattern of accents! They just do it. Figuring out how this is possible is one of the puzzles that make linguistics fun.

  All languages, whethe
r spoken in the deserts of Australia, the streets of Los Angeles, or the jungles of Brazil, use intonation. But many languages use pitch in another way as well. Although English uses pitch to change meaning in sentences, it does not use pitch to change the meanings of words, with some exceptions that can help us to understand what goes on in a tonal language like Pirahã or Chinese.

  Consider what differentiates pairs of nouns and verbs like CONtract (noun) versus conTRACT (verb), PERmit (noun) versus perMIT (verb), and CONstruct (noun) versus conSTRUCT (verb). In these pairs, the noun has higher pitch on the first syllable, while the verb has higher pitch on the second syllable.

  But while English uses pitch to distinguish meanings of only a few pairs of words, in tone languages, every syllable or vowel or word bears a distinctive pitch, called a tone.

  I first learned this distinction, as I learned many things about the language, by committing an egregious error. Kóhoi and I were working on some terms that I thought might be needed for Bible translation, among other things.

  I asked him, “When you like someone very much, what do you call him?”

  “Bagiái,” Kóhoi responded.

  I tried putting the word to use right away. “You are my bágiái,” I said, smiling.

  “No!” he responded, laughing.

  “What,” I asked, “you don’t like me?”

  “I do like you,” he said, giggling a bit. “I like you. You are my bagiái.” But there are bágiái and we don’t like them,” he clarified.

  To help me get what he was saying, Kóhoi whistled the words for me, slowly. Then for the first time I heard the difference! The word for friend is bagiái, with a single high tone on the last a: “ba-gi-Ai.” But the word for enemy has two high tones, one on each a: “bA-gi-Ai.” That little difference is what separates friend from enemy in the Pirahã language. The words are related to the Pirahãs because bagiái (friend) means literally “to be touching”—someone you touch affectionately—and bágiái (enemy) means “to cause to come together.” Culturally, though, bágiái has an idiomatic meaning—an enemy is someone who causes things that are not his own to come together. Idioms like this rely on more than the literal meaning of the words, which can in fact be irrelevant, like the English expression kick the bucket, which means “to die,” completely disconnected from the literal meaning of the individual words of the phrase.

  Clearly, I had to write the tones as part of the language. So I adopted a fairly common linguistic convention and used the acute accent to mark the high tone. When there is no mark above a vowel, the vowel has a low tone.

  Here is another set of Pirahã words, each one distinguished from the others by the relative pitch of the vowels:

  xaoóí (aoOI) “skin”

  xaoói (aoOi) “foreigner”

  xáoói (AoOi) “ear”

  xaóoí (aOoI) “Brazil nut shell”

  Because Pirahã uses pitch so widely, it has communication channel options that most European languages lack. I call these channels of discourse, following the pioneering work of sociolinguist Dell Hymes.There are five such channels in Pirahã, each having a unique cultural function. These are whistle speech, hum speech, musical speech, yell speech, and normal speech—that is, speech using consonants and vowels.

  To know the Pirahãs, one has to know about these channels and their functions. I had heard of these before going to the Pirahãs. And I knew that other languages had similar modes of expression (such as the drum languages found in Africa or the whistle speech of the Canary Islands). But when I heard an example in Pirahã for the first time, it was quite a novel experience for me.

  This happened one afternoon after I had set out some old National Geographic magazines for the Pirahãs to thumb through. They love pictures of animals and peoples, whether from the Amazon or other parts of the world. Xiooitaóhoagí (i-owi-taO-hoa-gI) sat on the floor, looking through the magazine, with her baby suckling at her breast. Her legs straight out in front of her, dress pulled down to her knees, in the normal Pirahã manner, she was humming rhythmically to the child on her lap as he nursed energetically. I watched for a bit before I realized that what she was humming was a description of the whale and Eskimos whose pictures she was examining. The boy would look away from her breast to the picture from time to time, and she would point and hum louder.

  Like all true communication channels, hum speech can “say” anything that can be said with consonants and vowels. But also like the other channels, it has a specific set of functions. Hum speech is used to disguise either what one is saying or one’s identity. It does this because even for a native Pirahã who is not paying close attention, it is hard to follow. And hum speech is conducted at very low volume. So it is also used for privacy, like our whispering. (The Pirahãs don’t whisper, they hum instead. I wondered about this for a while until German linguist Manfred Krifka reminded me of the obvious reason for it. In whispering, the vocal cords are unable to produce different tones, so Pirahã speech would be rendered unintelligible.) Hum speech is also used to talk when one’s mouth is full. Finally, it is frequently used by mothers when talking to their children.

  Yell speech is the use of the vowel a or, occasionally, the original vowels of the words being spoken and one of the two consonants k or x (glottal stop), to yell the musical form of the speech, that is, its tone, syllables, and stress. Yell speech is commonly used on rainy days when the rain and thunder are loud. It is used to communicate with Pirahã at long distances. It is as loud as a yell, but without consonants. It can occasionally be in falsetto.

  Koobio lives at Xagiopai, a seven-day canoe trip upriver from Posto Novo. One rainy day when I was there visiting, Koobio was across the river at his father Toitoi’s house. His wife, Xiáisoxái, was back at Koobio’s, about to cross to Toitoi’s side of the river. Koobio began yelling in this fashion.

  “Ká, Kaáakakáa, kaákaá.”

  In normal Pirahã speech this would be “Kó Xiáisoxái. Baósaí” (Hey Xiáisoxái. Cloth).

  Surprisingly, though the rain muffled most sounds, this speech carried impressively. Shortly thereafter we heard Xiáisoxái yell back, “OK, I’ll bring your shirt when I come.”

  And then there is musical speech, one of the two communication channels that the Pirahãs have special names for. They refer to musical speech as “jaw going” or “jaw leaving.” It is produced by exaggerating the relative pitch differences between high tone and low tone and changing the rhythm of words and phrases to produce something like a melody. This channel has perhaps the most interesting array of functions in the language. It is used to communicate important new information. It is used to communicate with spirits (and often it is used by the kaoáíbógí, or spirits, themselves). But it is mainly used when the people are dancing. Interestingly, though I have no explanation for this fact, when I ask Pirahã to repeat something in musical speech, women produce it less self-consciously than do the men.

  Whistle speech, which the Pirahãs refer to as talking with a “sour mouth” or a “puckered mouth”—the same description they use of the mouth when sucking a lemon—is used only by men. For some reason, this restriction to men is true of most other languages with whistle speech. It is used to communicate while hunting and in aggressive play between boys.

  My first intense contact with whistle speech came one day when the Pirahãs had given me permission to go hunting with them. After we’d been walking for about an hour, they decided that they weren’t seeing any game because I, with my clunking canteens and machete and congenital clumsiness, was making too much noise.

  “You stay here and we will be back for you later,” Xaikáibaí said gently but firmly.

  I watched the men leave me. I was standing by a large tree. I had no idea where I was or when they would come back. The jungle was dark from the shade of the climax forest trees. There were mosquitoes buzzing around me. I took out my machete in case an animal came prowling around. I wondered if the Pirahãs would ever be back to get me. (If they ha
dn’t returned, my skeleton would likely be there now.)

  As I tried to make the best of my solitary confinement, I heard the men whistling to one another. They were saying, “I’ll go over there; you go that way,” and other such hunting talk. But clearly they were communicating. It was fascinating because it sounded so different from anything I had heard before. The whistles carried long and clear in the jungle. I could immediately see the importance and usefulness of this channel, which I guessed would also be much less likely to scare away game than the lower frequencies of the men’s normal voices.

  These channels show how culture can influence language. If I didn’t know about the channels of discourse, I would not know the culturally appropriate way to communicate the different types of information that each of these channels is used for. A complete description of Pirahã culture has to include a discussion of how one communicates spiritual information, intimate information, and so on. The function of the channels is cultural. That is, the linguistic facts, the small number of phonemes and the amount of free variation between consonants that almost drove me crazy at the beginning of my work can’t be explained without the cultural information.

  Simply put, Pirahã may have so few sounds because it doesn’t need any more. The importance it places on these different channels makes consonants and vowels less important for the Pirahãs than they are for English, French, Navajo, Hausa, Vietnamese, and other languages. This challenges modern theories of language because these don’t expect culture to enter into the sound structure.