Fieldwork by phone

The telephone has a deal of history as a device for collecting data on languages. For example, the English language Switchboard Corpus was collected from telephone conversations in 1990-91 and, according to the manual:

“is a corpus of spontaneous conversations … [c]ollected at Texas Instruments with funding by DARPA, … [and] includes about 2430 conversations averaging 6 minutes in length; in other terms, over 240 hours of recorded speech, and about 3 million words of text, spoken by over 500 speakers of both sexes from every major dialect of American English”

Quite a number of researchers working on minority and endangered languages have also used phones to make calls from their offices or homes to their consultants in the field to collect data and/or check data and analyses. I recall Frank Wordick, author of The Yindjibarndi language [published 1977 by Pacific Linguistics, C-71 — for more on Yindjibarndi go here], saying in the mid-1970’s that he spent many hours calling his main consultants in Roebourne when he was back in Canberra following fieldwork in order to check aspects of his data. He even suggested he found it easier to distinguish retroflexes on the phone compared to face-to-face.

More recently, mobile phones (or ‘cell phones’) have become ubiquitous around the world, including in remote communities where other forms of communication, such as internet access, are restricted. They now provide a ready and almost instant way to contact speakers and to carry out fieldwork. I heard of several nice examples recently.

Lameen Soaug is a PhD student at SOAS who has recently returned from periods of fieldwork in Tabelbala, south-west Algeria where he worked on Kwarandzyey, a Songhai language (also called Korandje), and in Siwa, western Egypt where he worked on the Berber language Siwi. He has recently blogged as follows:

“Mobile phones have become ubiquitous even in such far-flung corners of the Sahara as Tabelbala and Siwa, used even by illiterate people — making it possible to keep asking people about the language well after you’ve gotten back to the university. So over the past months of fieldwork my phone has accumulated quite a lot of numbers, which I backed up to my computer today. The final count? At least 84 phone numbers from Tabelbala and 43 from Siwa. To put this in perspective, there are only about 3,000 Kwarandzyey speakers, so I can call something like 3% of the population.”

Perhaps the most spectacular example I know of comes from Marina Chumakina, a research fellow in the Surrey Morphology Group, who has been regularly using text messaging in her work on Archi, a Daghestanian language, for the past four years. Marina sends texts in Archi and Russian to her main consultant Bulbul Musaeva; if they are correct she either writes back правильно ‘correct’ or можно ‘allowed’, or war (imperfective of the Archi verb bos ‘say, speak, talk’ which means something like ‘people say that’). Occasionally she just replies ок ‘ok’.

If there are errors, Bulbul edits Marina’s text message, correcting any grammatical mistakes and explaining misunderstandings. Since Bulbul uses a phone with a Russian keyboard the corrections show up as Cyrillic characters when Marina gets the text message back, thus enabling her to easily see the relevant changes. She has now collected a corpus of over 150 sentences on her current phone (and has more on her old phone) using such a text messaging approach.

Here is an example that Marina emailed me concerning the handout for the Archi language tutorial that she, Grev Corbett and Dunstan Brown presented at last week’s LAGB conference:

“For example, I was trying to verify example 48 in our Archi tutorial handout, and sent the following text to Bulbul:

Privet, Buli! Pat’i ebqw?nittub gwachi buxlne pravilno ili nužno erqw?nittub? ili nužno

to which the reply was: Pat’iкьаш ebqw?nittub gwachi buxlne. So, she corrected the case ending and deleted the second variant of the verb. But then I realised that this means ‘the dog, which got scared of Patimat, ran away’, so I sent her another text saying: ‘no I need it to mean that Patimat is scared but the dog ran away’. Then she returned to me my first text message with the ebqw?nittub deleted and Pat’i in the absolutive, as you see it on the handout we used for the tutorial.”

Marina also used text messaging extensively in preparing the Archi dictionary (the on-line version is here). Again, to quote her at length:

“I used text messaging a lot to collect new words for the dictionary. Bulbul and her brother-in-law, Jaqub Magomedov, would send me Archi words, sometimes with translation, sometimes without. I would collect them up and work with the new words during my next field trip to Daghestan, asking other people whether such words exist in Archi, what they mean, and asking for example sentences of their use.

Such a method has the advantage that the speakers were always on the hunt for new words and could send them to me immediately, without having to look for pen and paper (not easy to find in Daghestan, and easy to lose afterwards). The two of them made a bit of a competition out of it, so when the whole family would be sitting at the table and the grandmother used an old word, Bulbul and Jaqub would try to be the first to text me this word. Archi are extremely competitive. They keep asking me questions like ‘who gave you the most words that you didn’t know before?’ or ‘how do I score against the rest of the village in terms of the size of my vocabulary?’ and so on”

So, as I pointed out last year, there are many possible models for successful fieldwork, including Crocodile Dundee fieldwork and ‘non-remote’ fieldwork with diaspora communities.

Now we also have opportunities for gr8 fldwrk th@ uses txt & mobile phone calls.

Postscript

I discussed this topic yesterday with SOAS PhD student Stuart McGill who is working on Cicipu, a Niger-Congo language of northwest Nigeria. Stuart made the following comments about phone fieldwork:

“I’ve been getting some text messages from one of my Cicipu consultants Mohammed Musa, and have added them to my Toolbox database. You find out about native-speaker intuition with respect to word breaks and spelling, but not much more than you would from pen-and-paper. Of course I have to add the tone marks myself! (Sophie Salffner [another SOAS PhD student, working on Ukaan from southwest Nigeria – PKA] does the reverse I believe — she calls people up to add the tone marks to material that they have written down for her).

The possibility of texting in Cicipu was a real eye-opener for Mohammed. I really think he was amazed that the language could be used in such a modern form of communication. He showed my texts to his brother, an influential Imam, and he was just as taken aback — and keen to support the development of the language. So I guess the fact that we are texting in Cicipu raises the status of the language a little in some people’s eyes.

Some people have been inspired to learn to read solely because they want to navigate their phones and send and receive text messages (phone menus are now available in Hausa, Yoruba, and Igbo in Nigeria). It would be really nice to have a Cicipu menu system (and predictive texting) on phones. This ‘should’ be easy to do technically and would potentially have a big effect on how people view their language, although I don’t know easy mobile phone companies make it to do this. [ JHS note: see James McElvenny’s post on this]

Finally, in our community orthography presentation we offered the (Danish) ø character as one of the options for the ‘open o’ vowel, since it is easier to type on a mobile phone. I know this won’t be a popular choice with typologists though!”

---

Note: Many thanks to Stuart McGill for information on his use of mobile phones in Nigeria, and especially to Marina Chumakina for discussion and providing examples from her phone with explanations. Note also that Archi has complex phonology that includes a fortis-lenis contrast, pharyngealisation, glottalisation, and vowel length. Marina and her consultants have developed an ad hoc orthography for texting that uses only Roman letters (from Marina) or Cyrillic letters (from Bulbul and Jaqub).