The purpose of this document is to outline a simple method for figuring out how to transcribe the tones of a tonal language.
All the words you elicit should have the same lexical category (e.g., nouns). You can extend the technique to other lexical categories (e.g., verbs) later.
Note: The purpose of this restriction is to simplify the task. Different lexical categories might be associated with different tonal patterns.
To start off, elicit all the words in isolation.
Note: In some languages, you do not get the full set of tonal contrasts for words elicited in isolation. In that case, a context might be necessary.
All the words that you elicit should have the same number of syllables. A good starting point is to only elicit two syllable words of the form CVCV (no onset clusters, no coda consonants, no syllabic nasals).
Use words that you have already elicited and transcribed in previous work.
Note: Have a list of these already-transcribed words in front of you, and try not to spend time obsessing over segmental transcription. Rather, for this exercise focus on tonal transcription.
You should set up the blackboard (or Google Drive file) to have a number of columns. Each represents one tonal pattern.
Note: You may need three to five columns, depending on the tonal system of the language.
Each time that you transcribe a word, place it in the column that it matches for tone. After you have transcribed a number of words, you chart will look something like this:
Column 1 Column 2 Column 3
word-a word-b word-c
word-d word-e word-f
How do you match a word to a column? Suppose you have just transcribed word-g. Here are some matching techniques:
You can ask the consultant to say words back to back. So, you can ask them to say word-d followed by word-g.
You can ask the consultant to hum the words. They can hum word-d, followed by humming word-g. Humming highlights the pitch of a word, and makes it easier to for the linguist to perceive pitch.
You can ask the consultant to whistle the words. They can whistle word-d, followed by word-g. Whistling highlights the pitch of a word, and makes it easier for the linguist to perceive pitch.
After a while the consultant might be able to do the matching for you. They may be able to tell you that word-g matches word-d in tone (once they get the hang of the task).
Note: Don’t get too dependent on humming and whistling. These are nice techniques, but in fact humans can hear the tonal distinctions without these techniques, so you should be able to as well.
In addition to placing a word in a column, you should confirm that the word does not belong in a different column. So, suppose that word-a goes in Column 1. Make sure to compare word-a to the words in columns 2 and 3 as well, to confirm that it does not belong there.
Tonal patterns are the way in which tones are matched up with the syllables of the words you elicit. For example, a CVCV word might have H on the first syllable, and L on the second syllable. So, the tonal pattern would be H-L.
After you have transcribed a few words and put them into columns, you can start trying to identify the tonal patterns. You will eventually name each column with a tonal pattern. For example, you might come up with the following:
H-H L-L H-L
word-a word-b word-c
word-d word-e word-f
How do you know what tonal patterns to use when you label the columns?
One technique is to try to hum the various possible tonal patterns. Hum to yourself H-H, L-L, H-L and L-H. Then when the consultant says a word, ask which of the patterns sounds more similar to what the consultant said.
12. We usually write tones as follows when transcribing.
One useful technique is to look for tonal minimal pairs, that is, two words that differ only in their tonal pattern. Tonal minimal pairs can be very useful when establishing the number of tones and tonal patterns in a language and the influence of surrounding segments on tone (e.g., depressor consonants).
After getting a few dozen words to fill the CVCV chart, you can move on to other words shapes. For example, you could try CV or CVCVCV words.
Remember that all of your results are provisional hypotheses. When you come to understand the tonal system of the language better, you will have a clearer idea of how to transcribe the tones accurately. You need to start somewhere!