Crowd Sourcing: LIP discussion

Lauren Gawne recaps last week’s Linguistics in the Pub, a monthly informal gathering of linguists in Melbourne to discuss topical areas in our field.

The topic of crowd sourcing is one that relates to many industries, and almost all participants in this month’s LIP have had some experience of crowd-sourced projects from the perspective of being part of the crowd. In this discussion though, we looked at the topic of crowd-sourcing specifically within the domain of endangered language documentation. This topic has two related, but distinct, facets. The first is the sourcing of funding from the crowd and the second is the sourcing of labour. We discussed both topics, and the benefits and down-sides we saw arising from each.

We started the discussion with crowd-funding. We spent a bit of time considering what would be needed to run a crowd-funding campaign with a chance of success. From a logistics point of view, you need a platform (Indigogo, Kickstarter, Pozible etc. all have different requirements). Then you need to craft a pitch for a specific, concrete and not-too-ambitious project. In this regard it’s possible that down-stream language documentation work may be more successful targets (such as the publication of traditional stories, or the salvaging of decaying tapes for archiving). Then you need to create an introductory video, have different prizes for amounts given, and a sustained social media and PR effort for the 4-6 week duration of the project pitch. This involves skills or time many of us may not have, and may not be supported or condoned by the University system. And all of this is predicated on the campaign being a success and gaining funding, which does not always happen. PARADISEC has had a sponsorship page, tax-deductible gift status, a paypal account and bank account, and has received a few small donations, only one of which was large enough to cover the cost of digitising a small collection of tapes.

With that in mind, it’s also important that these projects are largely driven by the community they are designed to assist. We talked about the difference in representing a community or language to a funding body, and representing them to the general public. The need to include the community may limit the types of projects and language groups for whom this kind of model may work – but it does offer an exciting possibility for those languages that have expat wealthier diaspora communitiesfor example. More generally, when it comes to representing communities for these crowd-funded projects many fell back on popular old tropes, like ‘save a dying language‘ that we spend a lot of time trying to challenge. While it provides a dramatic hook for bringing in members of the general public, we wondered if it does the communities a disservice.

Some have suggested that seeking funding from the crowd means that government bodies can reduce their funding of language documentation work. The general feeling from participants in the discussion was that governments will do that regardless. Crowd-funding will only ever likely work for projects looking for several thousand dollars, and won’t ever replace the larger grants needed for full projects.

As the dollar-values we are talking about aren’t really that great, we discussed some of the other benefits of crowd-funding. One that we kept coming back to was the way that it allows a broader audience to engage with language documentation and the predicament of the world’s endangered languages. As researchers or those who work with endangered language communities we can take our access to these experiences for granted, but there are many people who are interested in the work we do.

This brought us to the other side of the crowd-sourcing possibilities – the sourcing of help, or participation. You may already have come across ‘citizen science’ projects such as Zooniverse or Fold It. These involve people giving a small amount of time to do discrete tasks which are individually not taxing, but accumulate into a significant contribution. Another example is the Trove corpus of Newspapers in Australia, where people using the corpus can correct any erroneous OCR text to enhance the quality of the corpus.

Such a model could be used for the transcription of fieldnotes, or the identification of content in orphaned recording collections. This again allows people to participate, and would possibly allow us to enrich existing collections in more meaningful ways than paid labour would ever make possible. Participants in the discussion also felt it was less difficult and less problematic to ask for time than to ask for donations. Less problematic, because this kind of work would mostly be done on existing collections, and less difficult as people are generally more willing to give a small amount of time.

The challenge in crowd-sourced participation is building the platform to encourage it. Projects need to be identified that will have a chance of success, and the idea needs to engage people in micro-bites of accomplishment. Finding the initial funding to set it up, and then the ongoing time to engage participation can be as challenging as trying to raise money from crowd-funding. The good news is that as more people gain access to networked technology, the language expertise of participants will be even more diverse, allowing for more exciting projects.

Crowd-sourcing may be a permanent disruption to the way we have traditionally sought funding and assistance, or it may turn out to be a blip in the digital ecology, but at this time it offers some exciting and difficult possibilities.

[As a side note, this will be the last LIP summary post I will write as I am off to Singapore next month. Thanks to all the LIP participants who have added their ideas and thoughts to these summaries. Ruth will continue to run LIP, and I am sure that the Melburnians will continue to share their discussions via this blog.]

Follow