Tuesday, February 10, 2009

Asylum-seeker language test under scrutiny

ABSTRACT: "Language tests in immigration contexts typically perform a gate-keeping role in decisions about whether an applicant should be granted residence or citizenship in a new country. In refugee contexts, so-called language tests or language analyses also play a gate-keeping role, but a more ambitious one; namely that of providing answers to questions concerning the genuineness or honesty of asylum seekers' claims about their origins, whether national, regional, or ethnic. That is, the way that an asylum seeker speaks in an interview with immigration officials is analysed or assessed to help in the determination of whether to accept this person's claims about their origins. It is this assessment of language that is the subject of this article, in which I will explain the methods used and then highlight some problems that have been addressed by linguists. The acronym LADO is used to refer to such “language analysis” used for the determination of origin, but it should be understood that much of the “language analysis” in this area appears quite superficial"

Thus starts Dr Diana Eades article in the latest issue of Language Assessment Quarterly, warning that it is anything but scientific. "[They] are based on several folk linguistic views about the way people should speak that aren't borne out by the research on how language works," says Eades. A common or 'folk' view is that there is a one-to-one correspondence between nationality and language, but this is not true.

First, with widespread migration, there are immigrant groups in most countries who speak different languages. For example, Turkish in Germany, Vietnamese in Australia and Urdu in Afghanistan.

Second, national borders do not always coincide with linguistic borders. For example, a number of languages are spoken by indigenous people in Afghanistan as well as in neighbouring countries, including Dari (Farsi), one of the two national languages. In cases we have considered, members of the Hazara ethnic community from Afghanistan speak a variety of Dari which is also spoken in neighbouring Pakistan and Iran; Urdu, which is the official language of Pakistan, is also spoken by some communities of speakers in Afghanistan.

Many people believe that a person's national or ethnic identity can be determined by the use of a particular word or the way it is pronounced. But this is not necessarily true, because of two factors: language spread and linguistic change. Eades says that there are lot of subtleties in the way language works that makes this job not as easy as it would seem. There is language morphing and language leakage, for example, which means that there is nothing really called an "authentic" language for a particular area.

Not to mention that only trained linguists can provide some insight into the probable origins of a person. Many commercial companies being hired by governments (read the Swedish Eqvator) to analyse language are offering it as some kind of scientific fix that is relatively easy. "The department just sends off a cassette recording and some money and then they get back one of these one and a half page reports." The companies often use the judgement of native speakers who don't have linguistic training and have a narrow view of how the asylum speaker should speak. "It's equivalent to saying that because someone used the US term 'elevator' instead of 'lift', they are not from Australia, That's the level that these reports are operating on."

Eqvator and its notorious "language analysis" have been discussed since 2002. Tim McNamara, Professor of Linguistics at the University of Melbourne, published a paper in 2003 called "Linguistic Identification in the Determination of Nationality", in which he revealed that Eqvator, a privatised offshoot of the Swedish Immigration Board, carried out the language analysis relied on by the Australian government in determining whether it owes protection obligations to asylum seekers who arrive here without documentation. The work is done by interpreters, often themselves former asylum seekers, and others with varying but undisclosed degrees of linguistic training. The work of Eqvator has been the subject of ongoing criticism by linguists in Scandinavia. Professor Ruth Schmidt from Oslo University pointed to a number of problems with two Eqvator ' language tests' which she examined in 1997. She found that neither of them contained any scientifically recorded data for pronunciation or grammatical features, nor did they contain an adequate description of the language situation in the country from which the speaker claimed to come.

If this had been a simple linguistic discussion, that would be fine. But refugees are being refoulled back home to death and torture, because they happened to pronounce one word in 15 minutes of an interview in a language perceived as not their own!

No comments: