Witness Claims Accent Marked Troops' Origins

Defence lawyers in the trial of Congolese opposition leader Jean-Pierre Bemba last week contested claims by a linguistics expert who stated that people in the Central African Republic, CAR, would have been able to identify Congolese nationals based on their accent.

William Samarin, a professor of linguistics and anthropology at the University of Toronto, Canada, was called by prosecutors to offer expert opinion on whether CAR nationals would have been able to recognise Bemba’s soldiers on the basis of the languages they spoke and their accent.

Numerous prosecution witnesses have testified that soldiers from Bemba’s Movement for the Liberation of Congo, MLC, were the perpetrators of the widespread rape, murder, and plunder in CAR between October 2002 and March 2003.

These witnesses have testified that soldiers who brutalised civilians spoke Lingala - a language native to the Democratic Republic of Congo, DRC - and not Sango, the main language spoken in the CAR capital Bangui.

In a report he presented to court, and which formed the basis of his testimony, Samarin explained that Sango belonged to the Oubangian languages spoken in the Niger-Congo region.

He also noted that Lingala is a Bantu language spoken in the area around the Oubangui river, at the border of CAR and the DRC, all the way down to South Africa.

Samarin acknowledged that languages within each of the two families – Bantu and Oubangui – had a great deal of homogeneity. However, the families themselves were very different because they had distinct pronunciation as well as grammatical differences.

The expert also asserted that, as the national languages of their respective countries, Lingala and Sango each unified people in a “speech community”. As such, MLC soldiers would have used Lingala as a symbol of solidarity among themselves.

However, the defence observed that Samarin’s report was based on research carried out in 1994.

Samarin countered that “nothing has changed since 1994” as far as linguistic dynamics in the area were concerned.

The witness also clarified that in addition to his personal knowledge and his 1994 studies, his report was also based on the conflict in Bangui, and taken from material and witness testimonies provided to him by the Office of the Prosecutor.

However, defence lawyer Nkwebe Liriss observed that, in the professor’s data where he represented levels of knowledge and familiarity with Lingala, he had classified two prosecution witnesses of Congolese origin with the lowest level of familiarity – the same as witnesses of CAR origin who testified that they were unable to identify Lingala.

“If someone comes from Kinshasa [the DRC capital] that I gave Level 1, then I got mixed up. I should not pretend that this is an immaculate report,” Samarin responded.

Nkwebe also pointed out that another witness of Congolese origin was classified as Level 5, representing the highest knowledge and understanding of Lingala.

However, Nkwebe said this particular witness only had partial knowledge of Lingala and yet had perfect knowledge of Swahili. He said this witness actually gave his statement in Swahili and not Lingala.

Nkwebe also observed that, when estimating the percentage of CAR citizens able to recognise and identify Lingala, the professor only used their geographical proximity to DRC.

The study Samarin carried out was along the river that separates DRC from CAR, and most of the witnesses who have testified were Bangui residents.

The defence contended that in Bangui, just across from the DRC, far more people would have been able to speak or identify Lingala than the country as a whole.

Some crimes were committed far from Bangui, where the defence suggested that knowledge of Lingala would be less likely.

Samarin conceded that it was indeed unclear as to whether his conclusion on the number of Central Africans able to identify Lingala applied to the whole country or just Bangui.

“I should have been more precise with respect to inland and river land, but I wasn’t,” Samarin said.

Samarin agreed that besides Lingala, the alleged MLC soldiers may have spoken Swahili and other languages.

He added, however, that even if people from CAR knew Swahili or other languages spoken by the Congolese troops, they would have identified the soldiers as non-CAR citizens because they would have sounded different.

Nkwebe asked the witness why Lingala was identified as the language of the aggressors if indeed Congolese soldiers could have spoken other languages.

“Some people would have recognised bits and pieces of [Lingala],” Samarin said. “The important thing is not that they spoke Lingala, but that they spoke like people from over there (DRC). Of course they may also have had other uniqueness.”

Bemba is on trial for failing to stop or punish his troops as they allegedly raped, murdered, and pillaged in CAR during the 2002–2003 conflict. He has denied the charges, claiming that numerous other troops, both national and foreign, present in CAR at the time could have committed the crimes.

Bemba’s soldiers were in the country to help the country’s then president Ange-Félix Patassé fight off a coup attempt.

Samarin testified that historically, speakers of Sango and Lingala had not been in contact with each other. As a result, the two languages had not influenced each other in such a way that words from one were used in the other.

He said he had no doubt that people from CAR would have been able to identify Lingala based on personal experience, as they would have come into contact with the language through radio broadcasts from DRC as well as through Congolese music, which citizens of CAR were familiar with.

“Imagine a Central African tells us that ‘I recognised Lingala because in the market there are people who speak Lingala.’ Would you question him or her?” prosecuting lawyer Jean-Jacques Badibanga asked.

“I would not question that at all,” the witness replied. He added that such a person would not necessarily have to speak Lingala.

Samarin explained that even with different levels of recognition, the person would just say the speaker was from elsewhere and rule out the possibility that they were from CAR.

The witness also said DRC was the only country neighbouring CAR where there were Bantu-speaking populations.

He added that while there were a few CAR ethnic groups that spoke Bantu-type languages, these could not be confused with Congolese nationals because they also spoke Sango, a language widely spoken in CAR.

Furthermore, he said that even if the perpetrators spoke French or Sango, a language often spoken in CAR, CAR nationals would still have been able to recognise them as foreigners based on their accents, voice, and general “feel of the language”.

The trial continues this week.

Wairagala Wakabi is an IWPR-trained journalist.