Friday, January 11, 2008

The Sounds of Language, Part 3

Now that we've considered the vowels and consonants, what else could possibly be phonemic? What else can make a difference in meaning?

Just tell me why I would want to subject you to that subject?
Children whose conduct is exemplary conduct themselves well at all times.
Do you buy your produce in stores that produce many happy customers?
If you contract a mental disease, are you allowed to enter into a contract?
I'd like to present you with a present.
You can compare and contrast items only if there is a meaningful contrast between them.
The gentlemen on a protest demonstration protest their innocence too loudly.
An invalid may vote, and his/her vote will not be invalid.
The striking garbage collectors refuse to pick up the refuse.

I hope you get the point. Changing the stress from one syllable to another has the power to change the meaning of the word. Thus, we see that in English, stress is definitely phonemic. On the other hand, Hungarian is a language in which the stress always falls on the first syllable, no matter how long the word. So in that language, stress is not phonemic at all. It cannot change the meaning of a word.

Many people study Spanish and are familiar with its accent marks. Those little marks occurring above vowels tell us that one syllable is stressed more strongly than the others in the same word. If you shift the stress from one syllable to another, you might wind up with a nonsense word; or you might you might change the word's meaning in some way. For example, "hablo" is stressed on the first syllable, and means "I speak." "habló", on the other hand, stresses the last syllable and means "he/she/you spoke." "hable" is a command, meaning "speak!" "hablé" is a past tense, meaning "I spoke." "ingles" (stress on the first syllable) means "intestines, guts," "inglés" (stress on the last syllable) means "English." Changing the stress has the power to change the tense and the subject of the verb simultaneously, or, in the latter example, the complete meaning of the word. Thus, we have demonstrated that stress is phonemic in Spanish.

Stress is called a suprasegmental feature in a language. If stress can be proven to be phonemic, we then need to label suprasegmental phonemes separately from segmental ones (i.e., the consonants and vowels).

Can anything else be phonemic?

As a matter of fact, yes. The TONE of a word. To some extent, tone in English can be phonemic, but not in the same way that it is phonemic in Chinese. In English, if the TONAL CONTOUR shifts, it can create a difference in how the sentence is to be interpreted. We'll look at tonal contours in another posting. For the sake of this one, we'll assume that tone in and of itself is not really phonemic in English. However, tone is very much a phonemic reality to be reckoned with in all the Chinese languages and in others such as Zulu and many Native American languages.

A Mandarin word can be pronounced with one of four different tones. Each tone alters the word's meaning so that the same segmental phonemes pronounced in the different tones can mean mother, hemp, horse and also be used to indicate that what was just said is a question. That word is /ma/. With the suprasegmental toneme marker placed above the vowel, it could be /mā/ (first, or high tone), /má/ (second, or rising tone), /mă/ (third or falling-rising tone), /mà/ (fourth, or falling tone). Ma (with no tone markings) is termed a neutral tone. You can see that mother, hemp, horse and a grammatical indicator of interrogation have absolutely nothing to do with one another. One more fascinating example in Mandarin is that /măi/ means "buy" and /mài/ means "sell." You can really get into trouble if you confuse those two! That's the power of tone in Mandarin!

Cantonese is even more complex with its nine tonemes.

Zulu's two tones seem to be relegated to the personal pronouns. If a verb is pronounced with one tone, then the subject is "you;" if with the contrasting tone, then the subject is he/she. So tone is much more limited in its use in that language.

Consonants, vowels, stress and tone - these are the relevant distinctive features that make up the phonology of the languages of the world. In a future posting, we will begin to see how phonemes, though devoid of meaning on their own, combine to form particles - pieces of words which linguists call MORPHEMES, and that these little pieces, though they don't exist in isolation, are of vital importance in making up the longer units, the units we call "words."

Feel free to comment, query or add to any point made in this and other postings on this blog.

4 comments:

J. K. Gayle said...

What are the differences between African tone languages and Asian?

Did you know that speakers of Vietnamese (a "monosyllabic language" in which each "word" is a syllable with one of five or six different tones, depending on the dialect) cannot tell what word is spoken when it is whispered?

Thanks for the post.

El Profe said...

Thanks for the question and the information. Our friend Kenneth Pike actually wrote a book entitled "TONE LANGUAGES, A technique for Determining the Number and Type of Pitch Contrasts in a Language, with Studies in Tonemic Substitution and Fusion" published by Ann Arbor - The University of Michigan Press. At the time I purchased it, in 1967 or thereabouts, it had already gone through seven printings. You can probably still find it. In one of his footnotes, Pike comments that the great guru of Zulu tonemics, C.M. Doke, wavered big time in trying to ascertain the number of tonemes in Zulu. At first he thought there were nine, but Pike comments that not all of them were contrastive. Later on, Doke settled on three. I myself encountered only two in my brief fieldwork in 1969-70. And they occurred only in the pronominal system, separating 2nd person singular verb forms from 3rd person singular. See if you can locate the Pike book. Believe me, it is EXHAUSTIVE! He places the entire concept of tonemics under a witheringly scrutinizing microscope, after which he performs a fascinating analysis of two Mexican Native languges: Mixteco and Mazateco.
On your other comment regarding Vietnamese, don't you think it likely that though tone is inaudible when whispering, that the segmental phonemes would provide enough of a context for the native listener to be able to understand most of what was being whispered? As I know you well know, there's a lot of redundancy built into language, and I'd think that the other phonological and semantic clues would be enough to determine meaning. There's a Vietnamese restaurant a block away from my house (Saigon Grill). There are a lot of Chinese and other East Asian speakers there, but I haven't actually met a native Vietnamese there yet. I'll see if one of their cooks is Vietnamese, and I'll ask him about this. See if you can do likewise.

J. K. Gayle said...

Can't wait to look through Pike's book that you review briefly here--it's in the library here I see.

It's been more than 2 decades but I remember reading John Goldsmith's work on Autosegmental and Metrical Phonology, impressive enough but for African tonal (or perhaps intonal) languages. (And I think there was a fine rebuttal article: "How Abstract is Abstract"). It was novel stuff back in the late 1980s; is it still around?

So that took me to looking at Vietnamese (or Saigonese rather, which I'm a native speaker of). Goldsmith's model didn't work so well. Yes, as you point out, there are "segmental phonemes" and suprasegmentals in Vietnamese. But I was interested then in the "psychological reality" of the tones. 1 tone per syllable; 1 syllable (more or less) per "word." No tone change or shift independent of the syllabic context.

Then, at the time, came across a study of Vietnamese people whispering single words to one another with no context. What the researcher wasn't looking for, but what I found in the data was this: there were sets of tones, 3 pairs actually, that were consistently mistaken for one another. Without the vocalization of the tone (i.e., in the whisper) and without the larger context of other words and other vocalized prosody, I could only hypothesize two possibilities for the consistent mistakes: 1) there was variation in syllable length per tone pair or/and 2) there was variation in openness of the syllables. In fact, "2" is the the case in Vietnamese; that is, of the 6 tones (in Northern lects of the speakers in the study) there's a certain pair of tones that tends to be of syllables closed by glottal stops; and a certain pair that tends to be closed otherwise; and another pair that tends to remain open. That began my empirical search for proof of the "psychological reality" of syllabic tones in Vietnamese. Subsequent interviews of (other) native speakers did verify that they at least think that the 6 (or 5) tones are tonemic. And the African languages Goldsmith was saying operate independently of the individual vowel or syllable seemed to me a much different kind of language. Did the work, wrote a paper, but never sent it in for publication anywhere. Went running after all kinds of other interests (Now I wonder if Pike addressed all that already and we just didn't bring him into the discussion).

Anonymous said...

How often do you update your blog? More of the same, please! :-)
http://transubstantiation.wordpress.com/
Zak