Collecting Word List Data

Data Collection Tools
Interviews
Observation
Questionnaires
Recorded Text Testing
Sentence Repetition Testing
Word Lists
Participatory Methods
Matched-Guise

Word Lists
Preparing a Word List
Collecting Word List Data
Analyzing Word List Data
Interpreting Word List Data
Tips from the Field

Introduction Intense, unpredictable and often stressful, survey trips also provide untold opportunities to experience the richness of cultures that often have little exposure to the outside world. To maximise the ability to collect word list data, keep a flexible schedule, an open mind and a pen at the ready at all times. As you read, you might find it useful to refer to the Field Guide Glossary.

In the text below, fundamental steps in the process are highlighted in bold blue.

1 Finding a Participant

Before the elicitation session, make contact with local leaders (village leader, teachers, religious leaders, etc.) so you can explain the purpose for doing the survey and ask for permission to gather word lists. The local leaders can help you find good word list participants.

To help the community you find good participants:

• Ask descriptive questions

“Who do you know that could tell me a good story in this language?”

“Who knows the really difficult (or ‘inside’) words in this language?”

“Who would be a good person to teach me this language?”

• Be ready to describe the ideal participant in the LWC or another language

if you are not quick to identify your criteria, you may have participants who meet the assumptions and expectations of the person you asked, but don't match your criteria. The community might think an elderly person is a good person to teach you their language since they know it well and are free. But some older people might have difficulty understanding what you really want and you might have difficulty understanding their speech.

Find out if there is more than one variety spoken in that village. If there is more than one, then make sure you know which variety your participants represent. You may end up needing to elicit more than one word list from a site where people say there is more than one variety.

It is a good idea to have a group of two or three participants. Because they can check with each other about words they provide and is more enjoyable for everyone, both factors which increase the accuracy of the data. Having more than three participants might be too many as the elicitation will take too much time.

The person or group that you work with to collect a word list will have a large effect on the data. To be certain that you will get the right lect, the intended words, and a good recording, choose your participants wisely. Use the screening questions you have prepared to make sure you have acceptable subjects. You can also use observation, before you even ask any of the screening questions.

Make sure that your participants meet the following criteria:

• Clear speech

No speech impediments (e.g. has all their teeth, no cleft palate, no lisp, etc.)To be kind, you might suggest that such a person help by suggesting words to someone else who has no speech impediment. It is possible for someone to speak his or her language well but indistinctly. The clearer the speaker's voice, the easier you will recognize the sounds that you need to transcribe. The audio recording of the word list will also be clearer.

• Available

If a student is recommended, make sure that he or she doesn't need to be in school at that time. Word lists take a long time to elicit, and you do not want to choose someone who may have to leave in the middle of the list, or who will lose patience or become overly bored.

• Understands what you need

If you find a person who quickly grasps the nature and purpose of the list you want to elicit, you will probably shorten the time it takes to get a word list and improve its quality.

• Respected in the community

Sometimes the people most willing to help you are at the social margins of the community, and this factor may hinder you from getting a reliable list. Other people may be unwilling to accept such a person's involvement or his or her general qualification to assist you. If the person providing the word list has a good standing, it may give your research more credibility, interest, and cooperation.

• A representative speaker

sometimes people in the community feel that a person doesn't quite speak the local lect the way they feel it is generally spoken. If there are idiosyncrasies in a person's speech, then it is best not to elicit a word list from that person.

• Appropriate

Find out if it is appropriate to work with or get a word list from a person of that age and gender.

2 Finding a Recording Location

Consider carefully where you are going to make your recording. You want a location that will be available to you for the duration of the data collection and which has suitable furniture, lighting and air quality to enable both you and your participant/s to be comfortable.

In addition, consider sources of noise that will interfere with your recording. Children playing, roosters, bird calls, traffic and even heavy rain can render your recording inaudible when you later listen back to it.

3 Elicitation Protocol

There are many things to keep in mind when eliciting a word list. Do not expect to be able to remember all of these things the first time. As you gain more experience, you will know what kinds of problems to watch out for.

Before getting started, relax the participants in whatever way is culturally appropriate. This includes showing proper respect in your words and body language. Be friendly and fun in an appropriate way. Explain what you are going to do so the participants know what to expect.

First, say the word in the language of elicitation and ask them for their word for that. If necessary, explain more specifically what concept you are after. You may need to draw a picture or act it out. Make sure to have clear definitions for ambiguous words. For example, if in your part of the world, there are often two words for “back” depending on whether you are talking about the upper or lower back, be sure to consistently ask for one or both of them. Do not just say “back”.

Ask for the word that they typically use in that village. They may give you a clear borrowing. It is OK to ask if they have a native word as well, but you do not want to dig for a native word that they do not use. Find out what word they use the most. If they have a native word but hardly use it then consider the borrowing to be the word they use for that item. In some situations, you may have to really be persistent. They might have another idea in mind of what you want and give you words from another, more prestigious, dialect. Also, if the language of elicitation is very close to their language, they might just give you back the same word, or the same word with their local pronunciation. If that really is the word they use, great, but you might need to ask to make sure. If there is disagreement between them about the right word, let them discuss it and decide which the best word is.

Here are some additional tips for eliciting a word:

• If there are terms that you know to be confusing, have a consistent way of explaining them. For example, you could act the word out, or use it in a sentence.
• Make sure you know if you have elicited a generic term (e.g. animal) or a specific term (e.g. cat).
• If they give you what seems to be more than one word, find out what each part means. If there seems to be morphology, try to ask about what the parts mean.
• If you have data from other locations already, and the word they give you does not sound like any of the other words, you can ask them if they have words that sound like those other words. In the end, you want the word they use the most in their village. But asking about other words might reveal that there was confusion in the original elicitation. For example, maybe when asking for the word for “month” some locations are giving their word for “month” but others are giving their word for “moon.” This might happen if the language of elicitation has similar words for “month” and “moon.” If you discover this, then you can make sure you are getting the same concept from all further locations.
• There may be politeness levels in the language, so you may not know which level of the language you are using. Try to elicit words in the style of the language that is used in everyday speech. This is important when eliciting pronouns, or any other terms that differ across styles.

3.1 Clarifying Probes

If the elicitation probe isn't understood, you may be faced with one of the following responses:

• “I'm not sure which word to give you”

perhaps the probe could refer to one of a number of things or actions, and the person or group needs to know what you mean. Ask what the difference is from their point of view, and then let them know which one is the one you want.

• “I thought you wanted that other thing”

you may or may not find out that a probe was misunderstood. Listen for hesitation in the person providing the gloss, or look for nonverbal cues among the other people present. Someone else may know the right gloss.

• “I don't know what you're asking for”

the person or group didn't understand at all what you wanted. You may have to describe the object or action, refine an earlier description, or perform the action.

• “I don't even know what to say”

a blank stare is the most challenging response. Ask the person or group if they understand, or ask them what they think you said or asked for.

Another situation in which you might need to explain the probe and make decisions about the gloss to record is when the range of meanings for the suggested gloss does not match the semantic range of the probe. Identify the meanings that aren't covered by the gloss and ask for the words that describe them. Ask if there are any other words that describe the same thing or action. Write and record all these glosses, but mark the one that fits the item best.

3.2 Clarifying Data

Determining when you might have elicited a phrase instead of a word, and getting the right word instead, can be difficult. Length of the utterance is one clue. An unusually long utterance, as compared to other words on the word list, is always suspect. Do you recognize any part of the “word” as being similar to another word you've elicited? If so, you may have a phrase.

If you think you may have been given a phrase, try saying part of the utterance back to the assistant and ask if it means anything. If he or she looks confused, don't push it. Also, think back to how you elicited the word. Did you give a long explanation? Perhaps part of your explanation is in the “word” you elicited. For example, someone once elicited “that over there” instead of “that” (distal) because she explained it by literally saying in the LWC “that over there.”

Sometimes, the best help will be to look for possible synonyms. For example, when the participant hears something similar to “gate” in a related language or dialect, he or she will probably realize that a related word in his/her own dialect is better than the previously given “the door of the compound”.

Don't spend too much time agonizing over this problem. It will be equally agonizing for your participant, if not more so, and you definitely don't want to wear them out. It's enough to do a quick check of any suspect items.

Sometimes there is no equivalent for the item you are eliciting. In these cases, you will need to understand the intended meaning of the item, to know all the meanings that the probe can have, and to be familiar with synonyms and alternative ways of describing the item. Review the semantic range of the probe. Identify the ways in which the meaning of the probe and meaning of the suggested gloss don't match. Ask for the word (or words) that fills in the gap (or gaps), elicit synonyms with the same or similar range of meaning, and then choose the gloss that fits the item best.

The best equivalent may be unclear because two participants (or the members of the group) have provided different glosses and you aren’t confident that one is a better choice than the other. You might receive conflicting reports or claims about the meaning or “purity” of the words. If you still have access to either person (or the group), ask them to make sentences with each word and explain what the difference is. It may also be acceptable to seek another opinion – not to elicit the word again, but to get an opinion on the meaning and usage of the glosses and to help you choose one. If you are eliciting a list from the group, write down all the word forms that they provide, but mark the one that they choose as best. Keep in mind that they might have chosen that word because it is the most prestigious one or the most used form. Watch for one person to develop as the spokesperson and consider him or her as the final authority.

3.3 Dealing with Homophones

Beware of homophones in the language you are using to elicit the words. These can mean that the word you are trying to elicit will be ambiguous to the participant because it could be one of two or more meanings and results in transcriptions of words which are totally unrelated to the ones you want. If you are aware of possible homophones ahead of time, you can plan how to execute your clarification in advance. Homophones should have been identified and addressed during pilot testing, but in case they weren't, spend some time on them as soon as you discover a problem. Standardizing the way a confusing probe is clarified will possibly save you some time and frustration and will keep you from eliciting the wrong word. In addition to homophones, it may also be helpful to search for words that sound similar to other words. Remember that the language assistants are also usually second language speakers of the LWC, and even if your pronunciation is perfect, they may have trouble hearing slight differences

3.4 Consistency in Structural Form

Be consistent in eliciting with regard to structural form. Where consistency is a problem make a note of the difference. For example, you are eliciting verbs in a masculine form, but “give birth” must be in the feminine.

Always take the most generic term for an item on the word list. If no general term for a word is available, then use the name of the most common variety. For example, if no generic word for monkey exists in the language, and langurs are the most common variety of simian in the area, then take the word for langur. Make a note in parentheses that the word means langur, and not monkey. If there is no generic name for the domestic animal, elicit the form used for mature female animals (e.g., cow, not bull or calf). If adjectives are marked for gender, record the masculine form. In the case of verbs, two forms should be elicited. One of these should be the third person masculine form of the simple past tense (e.g., ‘he ran’), or its closest equivalent in the language. The second form may be the second person singular informal imperative, such as is used in giving a command to a child (e.g., ‘(you) go!’). If the language has more than one level of formality, elicit the command that is used to children.

3.5 Transcribing Data

Once the participant has decided on the right word, ask them to say it so you can transcribe it using IPA symbols. Refer to an IPA chart, if necessary, while you are transcribing. Look at the participant while they say the word. Pay special attention to lip rounding and point of articulation. Transcribe what you hear, repeating the word back to the participant if necessary. Do not ask the participant to repeat the word more than three or four times as this can become tedious. Trust your ears and go on. Better yet, video the participant saying the word. A phonetician will be able to help you even more if they can see the speaker’s face. Make sure to get permission before videoing an participant.

Here are some additional tips for transcribing words:

• Make notes in the margins of your word list book if there is anything about the word you want to remember. You can make additional notes that seem relevant (such as rounded, unrounded, open, falling, rising, breathy, dental, tense, creaky, etc).
• If you hear words that sound like minimal pairs, it may be helpful to ask the participant to say these two words together to isolate the differences. Record these minimal pairs, too.
• If there appears to be free variation, show both variants separated by a comma.
• You can check sounds that are unusual to you sometimes by asking the speaker to say given pairs of words that may clarify the sounds.
• Often the person you are speaking with will have linguistic ideas about his language. This will be helpful if they can give you words with the same tone, or ones that differ by one feature (vowels, consonants, times, rhyming, etc). However, sometimes a native speaker’s ideas may not be meaningful phonetically and too much of this type of discussion can take too much time.
• If you change your mind about a transcription, use a single line to cross out the incorrect transcription and write the corrected transcription next to it. Using a single line for cross-outs ensures that you can still read what you first wrote down. There are times when you realize that what you wrote first is actually better, or maybe you wrote the word for some other concept, and you want to know what you wrote. If there are some sounds that you just do not get, make a note of what that you think is significant (particularly if it has to do with the lips, mouth, or tongue position of the speaker) and use your audio recording to consult with a phonetician when you finish your fieldwork.

It is helpful to take a break during your elicitation session, so the participants and you do not get too tired. There can be a tension between taking breaks and finishing because an participant might leave and not come back. Balance this concern with the need for periodic breaks. Additionally, it is usually helpful for the participants to have some water to avoid getting a dry mouth.

4 Recording

The first time you elicit a word list, you may feel more comfortable transcribing and recording separately. A good way to do this is to elicit and transcribe 50-100 words at a time, then stop and record the words you've just elicited. When you are more experienced, you may want to transcribe and record at the same time. In this method, you would need to have the recording equipment set up all through the session. Once the word is agreed upon, you can transcribe the word while they are saying it three times. You could ask them to repeat it a few more times, if necessary, after you have stopped the recording.

To manage the recording process

• Find a quiet place for recording.
• Make sure the settings on your recording equipment are set correctly.
• Use fresh batteries for your recording equipment.

If the battery runs out while recording, some MiniDisc recorders will not save anything that you have recorded since the last time you pressed “Stop.” So if you have the habit of using “Pause” rather than “Stop” (because it is faster), make sure to have fresh batteries, and press “Stop” regularly to save the data.

• Place the microphone correctly.

If using a lapel microphone, make sure it is placed not too high (their chin will get in the way of the sound) and not too low (too far from mouth and you might record the sounds of their lunch being digested!). Placing it near their sternum is a good location.

• Wear headphones and listen in so you can make sure that the recording levels are correct.
• Test the recording set up to make sure everything is correct
• Record a “tag.”

This should include the date, your name, the language name (and dialect name if appropriate), the location (village name, district, and province), the informants’ names, and what is on the tape.

• Record all the words using the same informant.

If you do record more than one person, make sure to first have the whole list recorded from the same person. If you then additionally record some of the same words from other people, that will help you understand the variation in the language.

• Record word data consistently.

When recording a word, first you say the number of the word in English and then the word in English. Then you say the word in the language of elicitation and the informant says the word in his language three times. Saying the number as well as the word in English first helps a lot when you want to find where you are in the recording. Practice doing this a few times. Notice if the informant is saying the words unnaturally. For example, they might use a “list” intonation (e.g. putting a different stress or pitch on the last word). If the informant gives a different word when recording than when you were eliciting, make sure to take the time to sort out which word was the right one and what the other one meant.

• Reference data clearly

Make sure to label the tape or mini-disc using a permanent marker. Write the date, language name, location, and your name.

When you are all finished eliciting and recording, thank your informants and, if appropriate, pay them or give them a small gift.

Word Lists
Preparing a Word List
Collecting Word List Data
Analyzing Word List Data
Interpreting Word List Data
Tips from the Field

5 Checking a Word List

If possible, check the list with another informant from the same village. If you do this, you need to have a clear standard of what to do when you get conflicting information so that you can decide which person is right. If you check a word list, you will you probably have multiple forms at the analysis stage. You need to have a plan so you will know what to do with these forms.

If you double-check the word list:

• Do it in a way that does not embarrass the original informants.
• Use the informant screening questions to make sure the new informants are also representative of the same village.
• If the new informants give you a different word for a concept, try to find out what the difference is between the new word and the one you got from the first informant. Do they have the same meaning or different meanings? If they have the same meaning, which one is used more often in that village?
• If you make recordings of new words with a new informant, it is a good idea to record some of the old words as well. This will help you know if the new informant pronounces the same words in a different way.