Off topic: The DC Sniper - Translators & linguists in criminal matters
Thread poster: Bryan Crumpler
I\'m watching the news here in the states, and there\'s a lot of hype and blahdy blah about the serial sniper killing all those people in the Washington DC area.
Some new information has come to light about a letter of demands and threats that was written to the authorities. What those demands are, is beyond me. What\'s interesting, however, is just because the way the English was written, the authorities presumed: first, that the person was a non-native English speaker; second, that the writer was hispanic or a spanish speaking native. Their only description of the text, however, was that the English was \"broken.\"
Now, in communicating with Dutch people (and Americans/English people attempting Dutch), there\'s sort-of a common set of mistakes that are made, which are specific to the grammar of one\'s native language.
For example, in Dutch, an English speaker might make the mistake of saying \"Ik wil ruiken de thee\" when it should be \"Ik wil de thee ruiken\" (I want to smell the tea). This is a result of an English speaker not being used to Dutch word order when you use multiple verbs.
Another example would be a German person speaking English, who might make the mistake of saying \"I was yesterday at the store\", when it actually should be \"I was at the store yesterday.\" This is primarily because indicators for \"time\" usually come first in the word order of a predicate, whereas it\'s much more idiomatic to place time indicators at the end in English.
I even saw a post in the English (monolingual) forum about \"Hong Kong English\" which has a common trend of \"omitting\" things - particularly articles (\"a\", \"an\", and \"the\"). That would most likely arise due to the fact that articles aren\'t used in Japanese (or so I\'ve heard).
As linguists & translators (amateurs, professionals, learners and the like), I wanted to pose the question of how credible you think it is to link or trace grammar mistakes back to a particular source-culture or language? Has anyone ever studied this???
I can totally understand how _accents_ can link back to a language or country, but writing style????
It seems to me rather odd and not credible at all to make presumptions or establish the \"profile\" of a killer based on how they express themselves in written form.
I mean, not to throw in a race card, but because I write \"Yo, wassup man!\" should that indicate that I\'m black? Or because I write \"two dollar, please\" does that indicate I\'m Japanese? Or because I write \"I have 23 years old\" does that mean I\'m French????
Get the idea???
So... I\'m really curious what people think about studying the grammars of other languages and using that knowledge in criminal analyses to identify the cultural and language background of the author of a text written in their non-native tongue.
| || || |
| Contrastive analysis || Oct 24, 2002 |
There\'s a whole field of linguistics devoted to contrastive analysis, i.e. the comparison of the linguistic systems of two languages. I\'m sure linguists would be called in to analyze any written or oral information obtained by the police and if the samples were large enough they could probably pin down the author\'s language. But I have also found that some errors, such as \"I have 23 years\" are made by speakers of more than one language. Add Spanish to the French.
[ This Message was edited by: on 2002-10-24 17:03 ]
[ This Message was edited by: on 2002-10-24 17:05 ]
| | xxxAnneM
Local time: 19:00
Spanish to English
| The Mad Bomber Case || Oct 24, 2002 |
I\'m no expert but have an interest in criminology and profiling and a letter can tell a thousand tales, if you\'re trained obviously.
I\'ve always been fascinated by the case of the \'mad bomber\' over 60 years ago, when nobody had even heard of \'psycological profiling\'.
Summarised from a book:
The mad bomber planted 28 explosive devices around New York City over a period of 16 years, beginning in 1940. With the bombs came letters blaming an important electricity supply company.
In 1956, when a bomb injured 6 people, the police gave the letters to Dr. James Brussel (a pyschiatric consultant who became known as the \'Sherlock Holmes of the Couch\')
who, after studying the documents, told the detectives:
\"It\'s a man, Paranoic. Middle-aged. 45-50. Well-proportioned in build. He\'s single, a loner, perhaps living with an older female relative. He is neat, tidy and clean-shaven.Well-educated but of foreign extraction. He\'s a Slav. Probably lives in Bridgeport Connecticut (large Polish community). He has had a bad disease - possible heart trouble. When you catch him, he\'ll be wearing a double-breasted suit. Buttoned\".
They found him - George Metesky, living in Bridgeport Connecticut, well-proportioned, 51 years old, Polish extraction, unmarried, living with two older sisters and was wearing a double-breasted suit. Buttoned.
The book goes on to explain how Brussel came to each of these conclusions but I am sure there\'s loads of info. on the Internet.
I know your question was based more on \'linguistic reasoning\' but if this was being done in the \'50s, imagine what is known now.
| || || |
| This is very interesting ! || Oct 24, 2002 |
I believe it.
I think it has the same risks of judging from the face of somebody, but if the sample is large enough I think it really can tell lots of things.
I teach Spanish to French people (years of experience), and I can say where they are going to have trouble and what are the mistakes they are going to make. Sure! And when they don\'t make these mistakes I congratulate them, because I know the work they did. I also teach French to Spanish speaking people and it\'s the same case. But since I\'m teaching French to an English speaking person I\'m lost, because I don\'t know how it works. I think experience is the key.
Now, I\'d like you to tell me if I think in French or in Spanish when I write my English messages
[ This Message was edited by: on 2002-10-25 02:02 ]
| FBI Linguistic Analysts || Oct 25, 2002 |
I agree that this subject is fascinating. The FBI has Linguistic Analysts on staff whose job it is to analyze all forms of communication to try to profile suspects including their social background, education, whether they are male or female, etc. One example I read is that females tend to use more punctuation in written communication, such as using more exclamation marks and things like that. There is an interesting novel that touches on this subject written by Ken Follett called The Hammer of Eden.
Also, I do think that non-native English speakers from a particular country or culture do make a certain series of particular mistakes. I have been living in France and Spain for the past few years and, especially in Spain, have noted many mistakes that the Spanish make when they speak English but using Spanish grammar constructions or word order.
| Use of articles || Oct 25, 2002 |
I was going to say that the use itself of such basic elements as articles (and then: definite vs. indefinite) might reveal that we have to do with non-native speakers, when I read about the Hong Kong trend to omit articles in English and that, obvously, undermines my case. I would still argue, though, that what you can hypothesize based on the use of articles (given a sufficient sample) is that someone is not a native speaker of BE or AE. Which is of not much help since the writer can come from anywhere between Slavic countries and Hong Kong...
| Forensic Linguists || Oct 25, 2002 |
Sample abstracts from: http://web.bham.ac.uk/forensic/IAFL99/conf_abst.html
Prof. Hannes KNIFFKA
The University of Bonn,
Is it at all possible to analyze anonymous authorship (successfully) without comparison data of a suspect?
Mr. Krzysztof KREDENS
University of Lodz
This paper discusses the scientific tenability of expertise provided by linguists in the legal setting and in particular considers whether evidence based upon linguistic analysis can be utilised to convict. It is argued that the creation of a theoretical reference model based on the sociolinguistic concept of idiolect would greatly enhance the status of linguistic evidence in cases of disputed authorship.
| I believe in this as well || Oct 25, 2002 |
This is such an interesting topic!
I had my own little experience along those lines earlier this year when a Proz member posted a message in a forum.
This translator was commenting on the poorly written English message she had received from an agency. She did not specify any details as to where the author was from etc...
She pasted the message in the forum (without specifying any names or country etc) and asked for some feedback/comments.
I wrote to her privately, as when I saw the copy of the message, I immediately recognized the specific writing style.
It turns out that I had not only the right country of origin, but I also had guessed the name of the *person* and agency!
Obviously, I was able to figure it out simply because I had many dealings with the author of the mystery message (written in broken English). The writing style was definitively the essential clue in this case.
So, yes, I do believe that given enough samples and as Claudia mentioned \"dealing often enough with people in a certain linguistic situation\" (students in her case), you find out that there are some *traps* that people fall into consistently when writing in another language.
We all do it to a certain level
Have a wonderful weekend everyone!
| || || |
| | Terry Gilman
Local time: 19:00
German to English
| One suspect was from Brittany + aticom seminar on court interpreting || Oct 25, 2002 |
Thanks for starting this interesting thread.
What I saw on CNN didn\'t speculate on the nationality of the note writer, but I seem to remember that one suspect was a French marksman awol from a camp in Brittany (stuck in my mind as my brother-in-law is from Landerneau).
In a related vein, Aticom, a local translators and interpreters organization (I\'m not a member, just interested) will be holding a seminar on court interpreting on Nov 16 in Düsseldorf. I plan to go, work permitting, out of sheer interest in the relations between foreigners and German bureaucracy.
All the best,
| Writing samples are definitely exploitable lode || Nov 4, 2002 |
As Kim Metzger rightly notes, sample size is critical: if you have only one grammatical or lexicological error to go by, you will most often only be able to pin down a range of languages.
English is particularly rich because the spelling is so whacko (a distinction it shares with French): the same sound spells in so many different ways that second-language speakers also have ample opportunity for committing telltale spelling errors, e.g. somebody who writes \"responsability\" has fatally given her/himself away as French or Italian in the \"a\".
On the other hand, as a French-speaking native English speaker, I could write a piece that would pass for something written by a French(wo)man. But it would take time to read and reread and check for consistent errors. But in any event, the analyst would have tracked me down to one in 10,000 Americans with a better command of French, so if I\'m a suspect, the seeker has narrowed it down from 280,000,000 suspects to 28,000: nice start for an investigation, don\'t you think? (Admittedly, all figures are imagined here).
On the faking thread, it also helps to see if it feels like a hasty note or to know if the author had lots of time to write it. Coffee stains on the paper, for example, will communicate something, if only according to their pattern.
Even rules for indentation and layout are giveaways: in French, a writer with two words too many at the end of a line will right-justify them beneath the sentence in question, whereas an American will spill over into curl it up or down for a vertical read.
Even among English speakers, there a subtle distinctions. I once identified the author of a forex document as Australian because he wrote \"the Irish pound.\" Well, no economic or financial writer in Europe would write anything except \"(Irish) punt.\" There was the \'\"re\" giveaway on \"centre\", so...Aussie or Kiwi, right? Given the demographic odds, I bet on an Aussie.
There are all sorts of hints like this.
As Maria Reigger adds, you can distinguish gender -- I do it by the feel of the prose and am happy to learn she has indicators to quantify here.
I look forward to reading more on this thread.
| || || |