PDF versionPDF version
gna [Kaansa]
Word collection dates: 
Monday, 20 January, 2014 to Friday, 31 January, 2014
Total days: 
Obiré, Poni Province, Burkina Faso
Total participants: 
Avg daily participants: 
Avg hours/day: 
Avg participant groups: 
Avg Participants per group: 
Total domains treated: 
Total raw words collected: 
Local context: 
The workshop was held in the shade of three very large mango trees near the dispensary in the village of Obiré. The data entry was done in the SIL language program office about a kilometer away from the mango grove. A few participants from distant villages were lodged in Obiré, but others elected to travel by bicycle or motorcycle from as far away as 20km. Most traveled less than 10km each way.
% Words glossed: 
% Words entered: 
Software used: 
FLEx 8.0 stable candidate 2


The Kaansa Rapid Word Collection (RWC) workshop was a complete success in many ways, from the venue that was selected, to the number of participants involved, to the degree of preparedness on the part of the SIL team and logistics personnel involved, to the number of lexical elements collected, to the diversity of the Kaan population represented, to the level of devotion of the participants to the assigned tasks and goals, to the ability of the word-collection group members to work together productively and in apparent complete harmony for an entire two-week period, and the list could go on. Due to extensive and detailed preparations on the part of those organizing the workshop, a firm foundation was laid for what was to come. The enthusiastic participation in the training phase of 20-some individuals who would serve as coordinator, team leaders, scribes, glossers, typists, and record-keeper during the word-collection phase built solid walls on that foundation. And thus, it came as no surprise that, during the word-collection phase, the goal of collecting at least 10,000 words by the end of the second week was reached with relative ease.

In fact, there were some 50 people involved, including the Kaan king himself, and by the end of the 10th day of word collection, just over 11,500 words and expressions had been logged. Six word-collection groups each worked for 44 hours (264 hours total) over that time period to capture on paper the lexical knowledge of so many people. The glossers and typists all worked longer hours in their effort to add French equivalents to the words collected, and then get both Kaansa and French elements entered into the computer database. The dedication of these two groups of individuals kept the workshop from experiencing the same fate as so many others in the past—data on paper, but not in the computer. Even so, two additional hours were needed the week after the word-collection phase to complete the glossing, and five individuals all worked an additional day and a half to complete the data entry.

The location in the shade of a grove of mango trees near the dispensary in central Obiré, with a light breeze blowing most of the time, provided an atmosphere that enabled everyone to work well during the prescribed time from 8:00am to 2:00pm. And if there was a disappointing aspect to the workshop, it was only the fact that approximately one-fourth of the folders still remained untouched when time was up at 3:00pm on Friday, January 31st. However, in an effort to complete the remaining folders, the Kaansa team organized a follow-up week of word collection. (See the postscript to this report for details about that.)


The Kaansa workshop was organized according to the recommended allotment of time: 3 days of training, 10 days of word collection, and 5 days of data correction. However, rather than working the 7-8 hours per day prescribed by the best-practice formula, a shorter workday was adopted, due to concern that some of the (elderly) language experts would leave during the lunch hour and not return, or if they did return, that they would be too tired to accomplish much. (The people involved were, after all, not ones who are accustomed to working long hours on an intellectually oriented task.) Thus, lunch was delayed until 2:00pm, and word collection was suspended for the day at that point. While this schedule reduced the number of hours for collecting words, it did have the advantage of providing time each afternoon for the glossers and typists to work on catching up, attempting to complete their portion of the task for all the words that had been collected that day.

The prescribed six word-collection groups worked diligently through the folders that were made available to them, and these folders were processed as described in the best-practice formula, passing through the record-keeper’s hands at each step on the route from word-collection group to glossers to typists. The glossers found their task to be quite challenging, with many entries not having an obvious one- or two-word equivalent in French. But by working in teams of two or three individuals, they were able to almost keep pace with the word-collection teams’ output, though they had to stay on-site for another two hours each day to do so. The 3-5 typists, on the other hand, had to do a significant amount of “reprocessing” (i.e., correcting) of the data on the response sheets, so they were unable to keep pace with the daily production of the word-collection groups and glossers. This translated into an extra day and a half of data entry during the week that had been set aside for editing the data, thus reducing by 1/3 the time available for this latter task.

Unique attributes

In reflecting back after the workshop, there were a number of things that seemed to stand out as at least somewhat atypical. These are worthy of mention, so that others might consider incorporating them when possible.

Number of glossers

The pool of glossers was larger in this workshop than in the others I’ve been involved in before. This enabled them to work as teams, rather than individually. If this had no other benefic effect than to help maintain the morale of the glossers, it was a positive component of the workshop. I do believe that it did help them not to become “stuck” as often, however, and when they did, they had someone to dialogue with in hopes of making forward progress, instead of becoming discouraged by their apparent lack of ability to do the job they were assigned.

Depth of ‘participant pool’

The number of individuals who responded to the invitation to serve as language experts was greater than in any of the other workshops I’ve experienced. This was due, in part, to the fact that the organizers were uncertain that the primary invitees would really come, and so invited quite a few others as reserves. Not all of those primarily targeted came, but enough did that, together with all of the reserves who came, there were often 6-7 people per group, instead of the prescribed 4-5. This probably increased the number of words collected, but likely slowed progress through the folders at the same time, resulting in an excellent number of words collected per hour, but preventing the completion of all 126 folders during the two-week workshop.

When I first arrived in Obiré and saw a list of all who had been invited to attend, and in which capacity, the thing that stood out to me was the fact that, for every role, there was at least one individual mentioned as a potential substitute. Those who planned for this workshop planned very well, and it paid off in several ways, one of which was the relative ease with which a replacement was found for a team leader who quit coming after the first week.

Separation of typists from other participants

The word-collection portion of the workshop was held near the center of Obiré village, rather than at the translation and literacy office (located near the Protestant church in Obiré), so that the event would be seen as a community-wide event, and not one benefitting only the Kaan Christians. This choice of location contributed to the remarkable level of harmony and collaboration between Christians, Animists and Muslims that was evidenced throughout the entire workshop. It was also a very refreshing location physically, with the shade provided by the mango trees and the breeze that blew pretty consistently. But there was no electricity at that location, which would have made it difficult to have the typists do their work there. So it was decided that the typists would work a kilometer away from everyone else, in the SIL translation and literacy program office, where solar panels and batteries would enable the data entry to be done. Folders were then shuttled back and forth between the record-keeper and the typists, and this worked with minimal complication. The primary downside to having the personnel thus separated was the fact that the typists did not have easy access to the word-collection or glossing personnel when they had a question about something that they found on one of the response sheets. Thus we concluded that it is ideal to have everyone together in the same location, but if necessary, the typists can be in a separate location.

Folder selection limited

Rather than making all 126 folders available to word-collection group leaders right from the start, as was done in the first two workshops I led, we elected in this workshop to limit the choice to specific sections of the semantic domain hierarchy at any given time. We began by putting out the folders from sections 1, 2, and 5 and when all of those had been completed, we put out those from 6 and 7. After that, we made 3 and 4 available. A few folders from section 8 got into group leaders’ hands by mistake; otherwise, sections 8 and 9 remained untreated, along with several folders from sections 3 and 4.

Low variance in output among the word-collection groups

An unusual aspect of this particular workshop that continues to intrigue me is the relative parity of performance among the six word-collection groups. The number of words collected by each group during the 10-day workshop ranged from 1724 to 2136, with the intervening amounts being 1783, 1891, 1928, and 1965. Doing a little statistical analysis, this yields a mean of 1905, with a standard deviation of 145. The output of 4 of the 6 teams is within one standard deviation of the mean, with the two extremes lying not far beyond those boundaries. Though I don’t have the benefit of numerical data to back up my claim, I have the impression that there was much greater divergence in the production levels of the various word-collection groups in the previous two workshops I led than in this one. It will be interesting to see if these results remain as anomalous as they seem right now. I’d like to believe that the training that I gave the key personnel prior to the word-collection phase actually had an impact in equipping all groups to succeed at similar levels, but that is a hypothesis to be tested at this point. If subsequent workshops I facilitate manifest the same degree of parity of production between the various word-collection groups, such a premise would become increasingly credible.

Lack of interpersonal conflict and low-performance issues

Another unusual thing about the Kaansa workshop is the fact that, despite the random assignment of language experts to word-collection groups, we had neither of these issues that are common to RWC workshops:

  1. the need to substitute someone in a particular role because the person originally assigned to that role was struggling
  2. the need to reassign someone to a different word-collection group because s/he didn’t work well together with another particular person in the original group

We did have to substitute for one of the team leaders, when he simply quit coming to the word-collection sessions, but it was a relatively minor challenge to find someone to fill in for him.

Kaan culture eliminated certain potential problems

A possible explanation for the extraordinarily harmonious collaboration in the word-collection groups may lie in the fact that Kaan culture is by nature hierarchical, with people accustomed to taking orders from their superiors. The fact that they have a king and are brought up recognizing his authority might very well have meant that no one was bothered by being randomly assigned to work with a particular group of individuals; they just got on with what they were told to do. Different results are quite likely in an ethnic group without this same mentality.

Support and involvement of the Kaan king

In my estimation, the Kaan king’s attitude toward this word-collection workshop was a major factor in its success. The king not only supported the workshop in principle, but he spoke to the entire group of participants on the first day about the importance of collecting Kaansa words and of publishing a dictionary, and then he himself took part as a language expert as much as possible during the word-collection phase. His example, together with the hierarchical structure of the society, went far in motivating the participants to invest themselves significantly in the work. Without such support from the king, the results of the workshop might have been quite different.


As I dialogued with the workshop participants who spoke French (mostly the younger individuals), I was struck by the one comment that I heard time and time again: “I have heard many words during this workshop that I never knew existed in Kaansa.” Apparently, there are many lexical elements of the language that are not transmitted to other speakers of the language until they reach a particular age, or which are only used in very limited contexts that members of the younger generation are not involved in. These testimonials validated the workshop as an important tool for documenting the language and preserving this vocabulary for future generations.

Yao Jacob Farma, a member of the Kaansa translation and literacy team, had this to say, “The workshop was useful for me. I learned new words, for example kpɩ́nɛ which means ‘to be enslaved by or dependent on something.’ Previously I always said literally, ‘to be a slave of something.’ ” Yao was excited by the fact that he had had this opportunity to learn more about his mother tongue. Each of the others I talked with seemed similarly enthusiastic about having been part of the workshop.

Observations and recommendations

Help for team leaders with a limited level of bilingualism

A limited level of bilingualism in French was a handicap for many of the team leaders, and even for the glossers. There were numerous points at which team leaders needed help understanding the content of the questionnaire in order to translate it orally for their teams. Sometimes they would come to me as the consultant and ask for clarification, but many times they would simply do the best they could without outside help. When Thomas Blecke, a linguist from SIL Mali who came to observe during the second week of the word-collection phase, or I would sit with one group or another, the times that we were asked to provide clarification were much more numerous than the times that someone would leave his group to ask for help from one of the glossers or from me.

Since Thomas had worked for a number of years in Mali, he spoke Bambara, which is similar to Jula, the trade language in the area of Burkina Faso where Kaansa is spoken. Thus, he found that an explanation of the questionnaire content in Bambara was significantly more effective than even a paraphrase in French. He therefore began to routinely provide his explanations to the group leaders in Bambara, rather than in French. His presence was so much appreciated that, after a short while, the group he was sitting in on would not allow him to leave to sit with another group.

In light of these observations, Thomas and I concluded that it would be helpful to have any available “extra” people (consultant, consultant-in-training, coordinator, etc.) who understand the language of the questionnaire to spend as much time as possible sitting with the word-collection groups. They would not necessarily say anything, but would make it clear that they were there as resource people, and that team leaders should ask them for help whenever they had a question. Furthermore, for languages in Mali, western Burkina Faso, and northern Côte d’Ivoire, a translation of the questionnaire into Jula or Bambara would be a tremendous help. If RWC really catches on in that part of the world, consideration should be given to undertaking this non-trivial task.

Use of a pictorial dictionary

At the Kaansa workshop, we made use of a variety of books about trees, birds, mammals, reptiles, etc. But there was also a bilingual pictorial dictionary among the resources, which we belatedly discovered to be tremendously helpful for certain semantic domains. This book had pictures of a wide variety of things, with each image accompanied by lines pointing to various parts of the item, and the English and French terms indicated for those parts. The pictures, with the various parts illustrated, were helpful to the word-collection groups, while the French terms were useful to the glossers, who often struggled to find a precise French word for the Kaansa words they were trying to translate.

It would be a tremendous help in future workshops if this type of picture could be incorporated into the word-collection folders for the appropriate semantic domains. This might be in the form of photocopies of drawings of the style described, or it might simply be a reference to a page number in a resource book that is actually available at the workshop. It would be easy to obtain a book like the one that was available at the Kaansa workshop, and photocopy or scan pages from it, but there is the issue of copyright that would make incorporating scans in the questionnaire that come from a published book a difficult proposition.

Prioritization of folders

In retrospect, I would recommend that the folders from sections 5 and 6 be made available first, since the topics there are fairly concrete and there are some moderately productive domains that can serve as a good starting point (training ground). I would distribute folders 5.1 Household equipment, 5.3 Clothing, 5.4 Adornment, 5.5 Fire, 5.6 Cleaning, and 5.7 Sleep to the groups initially, as these are limited enough in scope that they should not take the word-collection groups too long to complete. Thus the glossers will not have to wait overly long for that first folder, and the consultant and coordinator will get prompt insight into any struggles a particular group might be having at the beginning of the process.

For the second set of folders to distribute, I recommend 6.2.2 Land preparation, 6.2.5 Harvest, 6.3 Animal husbandry, 6.4 Hunt and fish, 6.5 Working with buildings, and 6.7 Tool. The rationale for these choices is similar to that for the first set—keep the work concrete and of reasonable length at the beginning, then move on to the slightly more challenging folders after that.

Once a group has successfully navigated through a folder from each of the first two sets, I would give them free choice among the remaining folders from sections 5 and 6, plus all those from section 2. I would recommend not putting out for selection any other folders until all folders in sections 2, 5, and 6 have been claimed. As has always been recommended, the folders containing only the topmost level of a section (domains 2, 5 and 6) should be held back until all of the subdomains have been treated, to avoid the recording of a plethora of specific vocabulary in a very general domain.

Once sections 2, 5, and 6 have been completed, I would suggest making folders from sections 1, 3, and 7 available for selection by the team leaders, again withholding the top-level domains until all of the subdomains have been treated. Then finally, sections 4, 8, and 9 can be made available to the participants.

Treatment of specialized domains

In the RWC documentation on RapidWords.net, there are references to recruiting specialists in specific areas to serve as the language experts for domains relevant to their areas of expertise, but no details are given as to how best to go about doing that. In discussions with Stuart, Cathie, Thomas, and Anita toward the end of the Kaansa workshop, we decided that the best way to attempt to implement this would be to identify the specialized domains, separate them from the rest of the questionnaire, and put them in their own folders, based on who will need to treat them. Then designate one particular day (or part of a day) for working on these specialized domains. This seems better than trying to connect with individuals with specialized knowledge after the workshop is over, in view of expanding the lists of words collected for those domains during the RWC workshop.

On that particular day, invite those who have been recruited as specialists in particular domains to come and participate in one of the word-collection groups. The regular group leaders and scribes will continue to serve in those capacities, but with different group members on that particular day. When all of the specialized domains have been dealt with, the “specialists” can either continue working on general domains until the end of the day or they can be released to return to their other preoccupations and the general language experts can resume the work on the remaining semantic domains.

Therefore, we are recommending that future workshops plan ahead of time to identify the people who should be involved in the treatment of those particular domains, and a day assigned for those particular individuals to attend the workshop and work through the domains relevant to them. Team leaders, scribes, glossers, and typists would all work as normal on that day, but the language expert pool would consist of people with specific expertise, rather than the general masters of the language invited to participate on the other days of the workshop. Except for cases where an individual is in both the general language expert pool and in the pool of those with specific areas of expertise, the general language experts would be off-duty that particular day (perhaps the Friday of the first week of the word-collection phase).

The specific domains which should be dealt with by individuals with specific areas of expertise are still to be identified. That is a task that is on my to-do list to be drawn up sometime this month (March 2014).


The Kaansa RWC workshop was held in Obiré, Burkina Faso, January 15-February 7, and followed the best-practice formula in every regard except two: 1) the typists did their work about a kilometer away from where the word collection took place, and 2) the work days were shortened in order to discourage absenteeism and excessive fatigue. The participants worked together amazingly well and accomplished what they set out to do, namely, to gather a minimum of 10,000 words; in fact, they collected about 11,500. The young people were encouraged to see the value attributed to their language and culture through this workshop, and marveled at the fact that they had heard so many Kaansa words for the first time, as a result of this exercise. The only disappointment was the fact that not all semantic domains were able to be treated during the workshop. Some 500 domains remained, and the Kaansa translation and literacy team is working on a way to complete what was so auspiciously started in January 2014.


I received this report from Stuart Showalter, Kaansa team coordinator, in April, 2014:

The village team carried out four days of data collection in early March and was able to finish all the folders. They recorded another 3,500 words and expressions, bringing the grand total to 15,000 entries. With a few more days' work, they typed all the new data into Fieldworks as well. Because the team was well trained in January, and because the seminar went off well at that time, they were able to organize and carry out this second session entirely on their own, funded by the SIL Kaansa Project

Congratulations to the Kaan people for a job well done!