Embedded Files
Sound Sorting
Sound Sorting
When the word collection stage is complete, the next step is to sort cards by phonemes. The process described here starts by first studying vowels, then consonants, and lastly, tone.*
*Depending on the context, this ordering may need to be adjusted. For example, during a workshop on the Maꞌdi Urule language in Uganda , the functional load of tone was so high that participants found it necessary to first add tone markings (high, mid, and low) in order to sort vowels effectively.
Photographer Credit:
Zeke du Plessis
"This approach includes active involvement of native speakers in the analysis of the basic phonology. It was effective for helping the Naca people to understand the sound system of their language, not only the inventory of phonemes, but also the basic concepts of consonant, vowel, and tone."
-Shane Devereux & Jenita Van den Belt, Facilitators (Southeast Asia)
Overview: Data Analysis
Overview: Data Analysis
Goals:
Determine the number of vowel & consonant phonemes
Identify the basic tone system & any lexical tone
Methodology:
The SAME-DIFFERENT Test:
Grouping words by the feature (vowel, consonant, tone, etc.) being analyzed
Time Estimate:
Vowel System: 2-3 days
Consonant System: 1-2 days
Tone System: 2-3 days
Materials Needed:
Noun and verb cards from the word collection activity
Pencils
Rubber bands and/or resealable plastic bags for sorted words
Resources describing orthography options for the different vowels, consonants, and tone melodies discovered
Key Principles
Evaluate Nouns and Verbs Separately
Not all languages have the same phoneme set or tone melodies for both nouns and verbs. Nouns and verbs can also have different restrictions on which vowels may occur together within a word.
Begin analysis with word roots and obligatory affixes only. This reduces the complexity that can be caused by additional affixes.
Determining Phonemes
Determining Phonemes
PRL methodology establishes contrast, not by minimal pairs, but by proving that 2 sounds cannot be allophones. This is done by comparing sounds (and sound combinations) in the same structural position using the SAME-DIFFERENT Test. More detailed phonetic research comes afterward, where necessary. As the workshop progresses, the results of each step will help guide the next step in the process.
Rationale
Rationale
The SAME-DIFFERENT test, when done with precision, produces quality data that can serve as a foundation for additional research. It has several advantages over traditional methods:
1) Native speaker involvement results in a more accurate analysis than relying on an outside linguist's ear alone. Participants' comments will deepen the facilitator’s understanding of the language’s structure, resulting in more accurate data. For example, in a PRL workshop on the Naca language, participants noted a vowel contrast which had not been identified in research done before to the workshop.
2) The SAME-DIFFERENT Test determines contrast in situations where acoustic measurements cannot. Vowel phonemes which sound similar often partially overlap in their F1 and F2 formant plottings. When these plottings overlap, it is not possible to use them to clearly and accurately determine phonemes. (AFTER the language's phonemes have been established through other methods, acoustic measurements can be used to further study these contrasts and the acoustic overlap of certain vowels.)
3) The SAME-DIFFERENT Test more effectively determines contrast and complementary distribution than minimal pairs. Lists of minimal pairs often contain words and word forms which are not really comparable, as they may differ in word class and/or morphological structure. Meanwhile, the SAME-DIFFERENT Test prioritizes comparing sounds that occur in the equivalent syllable position of words in the same word class. In contrast to research results evidenced by minimal pairs alone, the results of the SAME-DIFFERENT Test show both syllable distribution patterns and the distribution of each phoneme.
4) In most languages, it is very difficult to find minimal pairs for all contrasting vowels. On the other hand, proof of contrast for any 2 sounds can be easily achieved by proving that it is impossible for them to be allophones.
Principles
Principles
Compare Identical Structures: Do not focus on assigning IPA symbols to individual sounds. Instead, examine sounds in context of their syllable position within word roots. The sounds being compared should occur in the same syllable position of the root, with nouns and verbs being treated separately.*
Focus on Identifying Contrast: The main goal is to determine which sounds are in contrast in order to establish a practical writing system. Avoid teaching articulatory phonetics, official IPA symbols, or analytical methods of doing phonology. The facilitating linguist can write the language's phonemes as official IPA symbols later on, if he or she would like to use the data for a linguistic publication.
Check Data: Checking and rechecking the sorted data is necessary for accuracy.
Note Minimal Pairs: While minimal pairs should not be used for proving contrast, they are still very useful for orthography guides and other resources, since they make the best examples for learning to read and write contrastive sounds. As minimal pairs come up during workshop activities (perhaps on sets of cards or through spontaneous examples), make note of these pairs for later use.
*In other words, the initial consonant of a prefix should not be compared with the initial consonant of a root, even if both consonants are word-initial. Likewise, a root-initial consonant that follows a prefix should not be compared with a root-medial consonant, even if both occur in the middle of the word. In general, avoid comparing sounds in a word root with sounds that result from combined morphemes.
Preparing the Data: Syllable Patterns
Preparing the Data: Syllable Patterns
In preparation for phoneme analysis, participants should sort the data by syllable patterns under the guidance of a facilitator.
Gather the words collected in Step 1, keeping nouns and verbs separate. By this point, compound words, loan words, and other word classes (adjectives, adverbs, etc.) should already have been removed from the data set.
Sort nouns into piles based on the syllable pattern of the word root, which may or may not have obligatory affixes. Be sure to take into account open (CV) and closed syllables (CVC) so that the words in every pile reflect the exact same syllable pattern, such as, CV-CV, CV-CVC, V-CV, etc.
For now, ignore any potential instances of long vowels.* Vowel qualities will be evaluated after phonemes have been established.
Sort verbs in the same way, beginning with the most common root structure. In Bantu languages, for example, the most common structure is often px-CVC-a, with occasional changes to the final vowel, depending on the vowel in the root.
When this initial sorting is done, the group can move on to researching the language's vowel system.
*If contrastive length exists, it may or may not have been written consistently, if at all. At this stage, any transcriptions suggesting a possible long vowel, for example CaaCa, should be kept in the same pile as their short vowel counterparts (for this example, the CaCa stack).
Quick Links:
Quick Links:
Page updated
Report abuse