Look up xenoglossy in Wiktionary, the free dictionary.
Xenoglossy (/ˈziːnɵɡlɒsi/ or /ˈzɛnɵɡlɒsi/), also written xenoglossia (/ziːnɵˈɡlɒsiə/ or /zɛnɵˈɡlɒsiə/), is the putative paranormal phenomenon in which a person is able to speak or write a language he or she could not have acquired by natural means. For example, a person who speaks Swedish fluently, but who is not a native Swedish speaker, has never studied Swedish, never been to a Swedish-speaking country, and never associated with Swedish speakers or had any other source of exposure to the Swedish language, would be said to exhibit xenoglossy. The existence of xenoglossy is not generally accepted by skeptic linguists such as Samarin (1976) and Thomason (1984, 1987, 1996). However, psychiatrist and paranormal researcher Ian Stevenson documented several cases that he considered authentic (Stevenson, 2001). The words derive from Greek ξένος (xenos), "foreigner" and γλῶσσα (glōssa), "tongue" or "language".
Scientific research into xenoglossy is quite rare. Ian Stevenson, a parapsychologist and psychiatrist at the University of Virginia, claimed there were just a handful of suggestive cases. These included two hypnotic regression cases where the subject could converse with people speaking the foreign language, instead of merely being able to recite foreign words. Sarah Thomason, a linguist at the University of Michigan, reanalysed these cases, concluding that "the linguistic evidence is too weak to provide support for the claims of xenoglossy".
A girl in India who lived entirely among Hindi-speaking people but was able to sing songs in Bengali, as identified by Professor P. Pal of Itachuna College in West Bengal, who studied the case after Professor Stevenson and transcribed some of the songs.
Uttara Huddar was a woman in India who normally spoke Marathi. While in the hospital undergoing psychiatric treatment, she began manifesting a personality called Sharada, who spoke in Bengali. Stevenson had recordings analysed by Bengali speakers, who disagreed among themselves about the subject's fluency. It cannot be ruled out that the subject may have learned Bengali earlier in life: both she and her father had a long-standing interest in Bengal, her home city had 1% native Bengali speakers, she had read Bengali novels in translation, and she herself had taken lessons in reading Bengali.
an American woman who presented the character of a Swedish farmer while under hypnosis conducted by her physician husband. Stevenson reported that the subject was able to converse in Swedish, albeit not fluently. However Thomason's reanalysis concluded that Jensen could not convincingly be claimed to speak Swedish; in the interview Stevenson studied in depth, though Jensen had a total vocabulary of about 100 words, only about 60 were used before interlocutors used them, and, as one of Stevenson's consultants pointed out, this reduced to 31 after eliminating cognates. Jensen also gave no complex sentences, mostly gave one or two word answers, and – as acknowledged by Stevenson – the subject's poor pronunciation was covered by correct spelling in the transcripts. Thomason mentions, however, that two of Stevenson's consultants praised Jensen's Swedish accent, and one claimed that only a native speaker could pronounce the word 'seven' correctly as Jensen does. Furthermore, she says that Stevenson's efforts to rule out fraud are convincing. Jensen's lack of understanding of Swedish was such that he answers ‘my wife’ to a question about what he would pay for some item at the market. Linguist William Samarin drew the same conclusion as Thomason.
an American woman named Dolores Jay who presented the life of a teenage girl in Germany while hypnotized by her Methodist minister husband. Stevenson reported that the subject was able to converse in German. Mrs. Jay did study a German dictionary at one point during the sessions, but Stevenson pointed out that she had already spontaneously produced 206 words before this event. Again Thomason's reanalysis, while acknowledging that the evidence against fraud was convincing, concluded that Gretchen could not converse in German. Her speech was largely the repetition of German questions with different intonation, or utterances of one or two words. Her "German vocabulary is minute, and her pronunciation is spotty". When asked what she had for breakfast, she answers ‘Bettzimmer’, which is a non-existent word made up of the two words for 'bed' and 'room'. Moreover she had some previous exposure to German in TV programmes and a "look at a German" book.
In 1931 a young girl from Blackpool, England began to speak in an ancient Egyptian dialect. She claimed to be under the influence of the personality of Babylonian princess and Pharaoh Amenhotep III's wife Telika-Ventiu, who supposedly lived about 3,300 years ago. Rosemary stated that she "hears" the Egyptian words clairaudiently and repeated them aloud. During more than a thousand language tests, the girl had spoken some 5,000 phrases and short sentences in the old Egyptian language. They were recorded phonetically and the first 800 of them were later identified and translated by an Egyptologist, Mr. Hulme. He claimed that Rosemary's speech substantially and consistently conformed to what Egyptologists know today of the ancient Egyptian tongue. Three books on the Rosemary case have been published and two gramophone discs of xenoglossy have been recorded. Wood claimed that Rosemary was never tested by independent witnesses and the claims were not submitted to independent scholarly inquiry.
In 1933 a 15-year-old well-educated Hungarian girl, Iris Farczády, who had dabbled extensively in mediumship, suddenly underwent a drastic personality change, claiming to be reborn Lucía, a 41-year-old Spanish working woman said by her to have died earlier that year. Transformed into 'Lucía', Iris spoke thereafter in fluent Spanish, a language she had apparently never learnt or had the opportunity to acquire, and could not understand any other language. Lucía has remained in control ever since and, aged 86 in 2005, still considered Iris to have been a different person, who ceased to exist in 1933. 
Proper scientific investigation of reports of xenoglossy is rare. More typical are press reports like that of Czech speedway rider Matěj Kůs from Pilsen, who, in September 2007 at the age of 18 reportedly awoke after a crash and was able to converse in perfect English. His ability did not last long and he was unable to remember anything from this episode. The press reports of his fluency in English are based entirely on the reports of his Czech team-mates. There is no record of his allegedly fluent speech or report by a skilled English speaker.
^Peter and Elizabeth Fenwick. Past Lives – an investigation into reincarnation memories. Berkley Books – Penguin Books, New York, 1999. ISBN 978-1-1012-0390-3 
^Mary Rose Barrington, Peter Mulacz and Titus Rivas, The case of Iris Farczády - A stolen life, Journal of the Society for Psychical Research, Volume69.2, Number 879, pp. 49--77, April 2005 http://www.txtxs.nl/artikel.asp?artid=738